Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.
No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.
Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.
We build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.
Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.
The Cognition Models team owns the models that turn video into structured understanding and reasoning: Pegasus, our video-language model, and Jockey Core, the reasoning LLM behind Jockey. In the model stack we sit between Perception Models (embeddings and retrieval) and the agent system — taking what's retrieved and producing structured understanding and the reasoning to act on it.
We focus on multimodal systems with high instruction-following capability and complex, hierarchically structured outputs. Our work spans training infrastructure from pre-training to RL, temporal segmentation and structured metadata extraction, large-scale inference and serving systems, data-curation and evaluation pipelines, and building Jockey Core. We ship products with real-world value rather than doing research in isolation, working as a goal-oriented, cross-functional team of ML researchers and engineers — using the most advanced compute in the world, including NVIDIA B300s, to accelerate the research-to-production cycle.
Pegasus is TwelveLabs' video-language model — it turns video into useful analysis by reasoning over visuals, speech, audio, and on-screen text. A key capability is Segment, our time-based metadata feature: instead of a broad question about a video, customers define the exact segment types they care about and the metadata fields they want back, and Pegasus finds the relevant start and end times and returns structured metadata for each segment — titles, summaries, topics, people, visual subjects, confidence, or domain-specific labels. This turns video into time-based, structured data that flows directly into search, archive, editing, compliance, or content-management workflows.
Drive research on Pegasus's harder problems such as temporal segmentation, multi-hour context, structured output generation, and training strategies from pre-training through RL, where the right approach requires deep judgment.
Design rigorous experiments and evaluation methods that produce clear signals on complex multimodal problems, including where ground truth is ambiguous.
Strengthen the team's research approach by helping reframe problems, sharpen hypotheses, and raise the bar for experimental rigor.
Work closely with ML Engineers to translate research advances into production, informing tradeoffs around architecture, serving, and system design.
Communicate research findings clearly and use them to inform technical direction across the team.
Explore and adopt AI-assisted development tools such as Claude, Gemini, and GPT to improve productivity across coding, experimentation, debugging, and documentation.
Significant research experience in one or more areas relevant to video understanding, such as multimodal LLMs, large-scale distributed training, temporal modeling, data-centric model development, computer vision, or vision-language systems, with demonstrated depth in at least one.
A track record of driving research on problems with significant technical ambiguity, demonstrated through projects, publications, or technical contributions.
Strong proficiency in Python and PyTorch.
Exceptional experimental judgment, including the ability to design evaluations for complex multimodal problems, run rigorous ablations, and draw clear conclusions from empirical results.
Strong communication skills and a track record of strengthening others' research through collaboration — helping formulate sharper hypotheses, identify more informative experiments, or reframe problems more tractably.
Experience working on multimodal systems involving video, vision, language, or structured output generation.
Experience improving model quality through data curation, evaluation design, or training data enhancements.
Experience with large-scale distributed training in high-performance GPU environments.
Experience translating research advances into production ML systems.
Experience defining research direction within a team or project.
MS, PhD, or equivalent practical experience in Machine Learning, Computer Science, or a related technical field.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Video lives in many existing locations today. For our largest customers, bringing the video intelligence to their data is the fastest way to deliver value. Our partnerships with platforms where that video already exists is a critical part of our business.
The Platform Integrations team owns that path: co-development, distribution, and integration partnerships with major clouds such as AWS, Azure, and GCP, and data platforms such as Databricks, Snowflake, and others. This role is also tip of the spear for future AI services integrations as the ecosystem evolves, the Platform Integrations team will be the one to bring TwelveLabs models to the platforms our customers need us most.
In this role, you will own the product roadmap for this ecosystem. You will decide which integrations are worth building, validate demand before engineering execution, and lead each integration from the first conversation through joint general availability. You will also work closely with the Head of Strategic Partnerships and VP of Revenue to design the strategy and build conviction on the roadmap.
The work requires balancing TwelveLabs' roadmap against each partner's incentives, constraints, and technical realities.
This is a staff-level individual contributor role with unusual leverage. One well-scoped integration with a major platform can unlock adoption for thousands of downstream.
The Platform Integrations team sits at the intersection of Product, Engineering, Science, and Go-to-Market.
We manage a portfolio of concurrent partner workstreams spanning early technical scoping through general availability and scale. The portfolio includes foundation model hosting, knowledge base and vector search integrations, media supply chain tooling, and infrastructure partnerships.
We work directly with partner product managers, solutions architects, and engineering teams to co-design integrations. Internally, we work with Engineering, Sales, Partnerships, Field Engineering, Security, Legal, Science, and Marketing to turn those integrations into pipeline, usage, and durable distribution.
Own the product roadmap for co-development, distribution, and integration partnerships across major cloud providers and data platforms.
Partner directly with product and engineering counterparts at platform companies to scope, prioritize, and ship joint integrations.
Identify the business and technical risks behind each proposed integration and design tests that validate partner and customer demand before engineering execution.
Translate ambiguous, multi-stakeholder partner requirements into clear product specifications that internal and partner engineering teams can execute.
Drive integrations end to end, including technical design, security and compliance review, GTM enablement, launch, and co-sell readiness.
Align Engineering, Sales, Partnerships, and Marketing on roadmap priorities, pricing, packaging, and joint go-to-market plans.
Track partner-sourced pipeline, usage, and adoption as the core success metrics for the integrations you own.
Hold the roadmap accountable to product and business outcomes rather than the number of integrations shipped.
Represent TwelveLabs in partner-facing technical and business conversations, with a clear understanding of both TwelveLabs' priorities and the partner's incentives and constraints.
7 to 10 years of product management experience, with meaningful work in partnerships, platform, ecosystem, or integration products. Demonstrated capability matters more than tenure.
Experience in a customer-facing solutions architecture or forward-deployed engineering role, with real depth in system architecture and/or machine learning.
Strong technical fluency in cloud infrastructure, system architecture, and API design. With an emphasis on GPU heavy integrations.
A track record working with major cloud or data platforms such as AWS, Azure, GCP, Databricks, Snowflake, or similar companies and their AI/ML services.
Clear and concise written communication across technical specifications, partner briefs, and roadmap documents.
Comfort operating in ambiguity, managing multiple partner relationships at different stages, and reprioritizing as new information arrives.
A platform and ecosystem mindset. You look for leverage and network effects rather than treating every request as an isolated feature.
A bachelor's degree or equivalent practical experience.
Experience with hyperscaler co-sell or partner programs such as AWS Partner Central and ACE, Azure Marketplace, or GCP Partner Advantage.
Familiarity with foundation models, vector or embedding-based data platforms, or AI/ML inference infrastructure.
Experience operating in a startup or high-growth environment with fast-shifting priorities.
Background in media, video, or content technology.
🤝 An open and inclusive culture and work environment
🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology
🏥 Full health, dental, and vision benefits
🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years
💪 Monthly wellness stipend
📚 Annual Learning & Development stipend to invest in your growth
🛂 VISA support where applicable
🚆 Transportation stipend (SF onsite only)
🍲 Daily lunch & dinner provided (SF onsite only)
","descriptionPlain":"WHO WE ARE\n\nVideo is 90% of the world's data. Most of it is invisible to machines.\nTwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.\n\n\nWe have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.\n\n\nWe are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!\n\n\n\n\nABOUT THE ROLE\n\nVideo lives in many existing locations today. For our largest customers, bringing the video intelligence to their data is the fastest way to deliver value. Our partnerships with platforms where that video already exists is a critical part of our business.\n\nThe Platform Integrations team owns that path: co-development, distribution, and integration partnerships with major clouds such as AWS, Azure, and GCP, and data platforms such as Databricks, Snowflake, and others. This role is also tip of the spear for future AI services integrations as the ecosystem evolves, the Platform Integrations team will be the one to bring TwelveLabs models to the platforms our customers need us most.\n\nIn this role, you will own the product roadmap for this ecosystem. You will decide which integrations are worth building, validate demand before engineering execution, and lead each integration from the first conversation through joint general availability. You will also work closely with the Head of Strategic Partnerships and VP of Revenue to design the strategy and build conviction on the roadmap.\n\nThe work requires balancing TwelveLabs' roadmap against each partner's incentives, constraints, and technical realities.\n\nThis is a staff-level individual contributor role with unusual leverage. One well-scoped integration with a major platform can unlock adoption for thousands of downstream.\n\n\n\n\nABOUT THE TEAM\n\nThe Platform Integrations team sits at the intersection of Product, Engineering, Science, and Go-to-Market.\n\nWe manage a portfolio of concurrent partner workstreams spanning early technical scoping through general availability and scale. The portfolio includes foundation model hosting, knowledge base and vector search integrations, media supply chain tooling, and infrastructure partnerships.\n\nWe work directly with partner product managers, solutions architects, and engineering teams to co-design integrations. Internally, we work with Engineering, Sales, Partnerships, Field Engineering, Security, Legal, Science, and Marketing to turn those integrations into pipeline, usage, and durable distribution.\n\n\n\n\nIN THIS ROLE, YOU WILL\n\n - Own the product roadmap for co-development, distribution, and integration partnerships across major cloud providers and data platforms.\n\n - Partner directly with product and engineering counterparts at platform companies to scope, prioritize, and ship joint integrations.\n\n - Identify the business and technical risks behind each proposed integration and design tests that validate partner and customer demand before engineering execution.\n\n - Translate ambiguous, multi-stakeholder partner requirements into clear product specifications that internal and partner engineering teams can execute.\n\n - Drive integrations end to end, including technical design, security and compliance review, GTM enablement, launch, and co-sell readiness.\n\n - Align Engineering, Sales, Partnerships, and Marketing on roadmap priorities, pricing, packaging, and joint go-to-market plans.\n\n - Track partner-sourced pipeline, usage, and adoption as the core success metrics for the integrations you own.\n\n - Hold the roadmap accountable to product and business outcomes rather than the number of integrations shipped.\n\n - Represent TwelveLabs in partner-facing technical and business conversations, with a clear understanding of both TwelveLabs' priorities and the partner's incentives and constraints.\n\n\n\n\n\nYOU MAY BE A GOOD FIT IF YOU HAVE\n\n - 7 to 10 years of product management experience, with meaningful work in partnerships, platform, ecosystem, or integration products. Demonstrated capability matters more than tenure.\n\n - Experience in a customer-facing solutions architecture or forward-deployed engineering role, with real depth in system architecture and/or machine learning.\n\n - Strong technical fluency in cloud infrastructure, system architecture, and API design. With an emphasis on GPU heavy integrations.\n\n - A track record working with major cloud or data platforms such as AWS, Azure, GCP, Databricks, Snowflake, or similar companies and their AI/ML services.\n\n - Clear and concise written communication across technical specifications, partner briefs, and roadmap documents.\n\n - Comfort operating in ambiguity, managing multiple partner relationships at different stages, and reprioritizing as new information arrives.\n\n - A platform and ecosystem mindset. You look for leverage and network effects rather than treating every request as an isolated feature.\n\n - A bachelor's degree or equivalent practical experience.\n\n\n\n\n\nPREFERRED QUALIFICATIONS\n\n - Experience with hyperscaler co-sell or partner programs such as AWS Partner Central and ACE, Azure Marketplace, or GCP Partner Advantage.\n\n - Familiarity with foundation models, vector or embedding-based data platforms, or AI/ML inference infrastructure.\n\n - Experience operating in a startup or high-growth environment with fast-shifting priorities.\n\n - Background in media, video, or content technology.\n\n\n\n\nBENEFITS AND PERKS\n\n🤝 An open and inclusive culture and work environment\n\n🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology\n\n🏥 Full health, dental, and vision benefits\n\n🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years\n\n💪 Monthly wellness stipend\n\n📚 Annual Learning & Development stipend to invest in your growth\n\n🛂 VISA support where applicable\n\n🚆 Transportation stipend (SF onsite only)\n\n🍲 Daily lunch & dinner provided (SF onsite only)"},{"id":"8ad28030-3654-4793-a7e1-25611e29fbd0","title":"Senior AI Engineer, Tools & Agents","department":"Tech","team":"Engineering ","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-07-20T22:23:13.166+00:00","isListed":true,"isRemote":true,"workplaceType":"Hybrid","address":{"postalAddress":{"addressRegion":"CA","addressCountry":"United States","addressLocality":"San Francisco"}},"jobUrl":"https://jobs.ashbyhq.com/twelve-labs/8ad28030-3654-4793-a7e1-25611e29fbd0","applyUrl":"https://jobs.ashbyhq.com/twelve-labs/8ad28030-3654-4793-a7e1-25611e29fbd0/application","descriptionHtml":"Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.
No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.
Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.
We build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.
Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.
We're hiring a senior AI engineer to own the integration layer that makes TwelveLabs' video AI accessible to the world.
Jockey is TwelveLabs' multimodal agent for video and image understanding. It's built on a knowledge store of ingested content, retrieval primitives like search, and an agent layer that orchestrates those primitives: planning, calling them, and reasoning over results to return grounded, cited answers. The engineer in this role owns the surfaces that make this system accessible to external developers and AI agents.
The engineer in this role decides how AI agents and developers connect to these capabilities, and you'll build the agent harness and the supporting infrastructure that makes that connection trustworthy at enterprise scale.
The center of gravity is our MCP server. You'll own it end-to-end: how agents discover and invoke our capabilities, what the tool interfaces look like, how failure modes are handled, and how the surface evolves as the agentic ecosystem does. Everything else, auth, metering, rate limiting, is the substrate that makes that surface something an enterprise customer will trust.
Location: San Francisco. Onsite or hybrid. No fully remote option.
The agent-facing integration surface. TwelveLabs' MCP server and agent-to-agent interfaces are yours to design and operate. You'll decide how AI agents discover, connect to, and build on Jockey's retrieval and reasoning capabilities, and you'll be the person who knows what breaks and why.
The auth and access substrate. OAuth, RBAC, and multi-tenant isolation for enterprise customers. This is the work that turns a powerful capability into something a large organization will put in production. You'll design it from scratch rather than configure someone else's framework.
The enterprise-readiness layer. Per-API-key usage tracking, rate limiting grounded in real system constraints, and the reliability work that makes the platform trustworthy under real load. You'll define how the platform behaves under pressure before customers find out the hard way.
The research partnership. You'll work directly with our research team to turn frontier video and image understanding capabilities into durable product surfaces. The loop from research to what customers can build is short, and you're a key part of closing it.
You've shipped real agentic or LLM-powered systems. Not demos, but production systems with real users and real failure modes you had to debug and fix. You can talk about what broke, why, and what you changed.
You understand how retrieval, reasoning, and agent orchestration fit together as a system. Jockey's architecture sits at that intersection, and the integration surfaces you build need to reflect how that system actually behaves under real conditions.
You've owned a developer-facing or external API end-to-end, with real decisions about contracts, versioning, and the experience of building on top of your surface.
You've built authentication and authorization in production, OAuth, OIDC, RBAC, or multi-tenant isolation, as a primary owner and not a consumer of someone else's framework.
You've thought seriously about enterprise-readiness: what it takes to go from a capability that works to a platform that a large customer will trust with their data and workflows.
You bring versatility across the full platform surface. We're not looking for a specialist in one of these areas. We're looking for someone who can own all of them and make good tradeoffs across them.
You've operated in a fast-moving environment before, a startup, a hypergrowth company, or an AI-first team where you shipped under ambiguity without heavy process support.
Experience with MCP, A2A protocols, or the broader agent ecosystem. You've contributed to or built on these surfaces and understand where they're headed.
Hypergrowth pedigree from a technically excellent startup.
🤝 An open and inclusive culture and work environment.
🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology.
🏥 Full health, dental, and vision benefits
✈️ Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years.
🛂 VISA support where applicable
","descriptionPlain":"WHO WE ARE\n\nVideo is 90% of the world's data. Most of it is invisible to machines.\nTwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.\n\n\nWe have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.\n\n\nWe are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!\n\n\n\n\nABOUT JOCKEY\n\nJockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.\n\nNo context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.\n\nBuilt for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.\n\nWe build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.\n\nDeep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.\n\n\n\n\nABOUT THE ROLE\n\nWe're hiring a senior AI engineer to own the integration layer that makes TwelveLabs' video AI accessible to the world.\n\nJockey is TwelveLabs' multimodal agent for video and image understanding. It's built on a knowledge store of ingested content, retrieval primitives like search, and an agent layer that orchestrates those primitives: planning, calling them, and reasoning over results to return grounded, cited answers. The engineer in this role owns the surfaces that make this system accessible to external developers and AI agents. \n\nLearn more about Jockey here https://www.twelvelabs.io/jockey. \n\nThe engineer in this role decides how AI agents and developers connect to these capabilities, and you'll build the agent harness and the supporting infrastructure that makes that connection trustworthy at enterprise scale.\n\nThe center of gravity is our MCP server. You'll own it end-to-end: how agents discover and invoke our capabilities, what the tool interfaces look like, how failure modes are handled, and how the surface evolves as the agentic ecosystem does. Everything else, auth, metering, rate limiting, is the substrate that makes that surface something an enterprise customer will trust.\n\nLocation: San Francisco. Onsite or hybrid. No fully remote option.\n\n\n\n\nIN THIS ROLE, YOU WILL OWN\n\n - The agent-facing integration surface. TwelveLabs' MCP server and agent-to-agent interfaces are yours to design and operate. You'll decide how AI agents discover, connect to, and build on Jockey's retrieval and reasoning capabilities, and you'll be the person who knows what breaks and why.\n\n - The auth and access substrate. OAuth, RBAC, and multi-tenant isolation for enterprise customers. This is the work that turns a powerful capability into something a large organization will put in production. You'll design it from scratch rather than configure someone else's framework.\n\n - The enterprise-readiness layer. Per-API-key usage tracking, rate limiting grounded in real system constraints, and the reliability work that makes the platform trustworthy under real load. You'll define how the platform behaves under pressure before customers find out the hard way.\n\n - The research partnership. You'll work directly with our research team to turn frontier video and image understanding capabilities into durable product surfaces. The loop from research to what customers can build is short, and you're a key part of closing it.\n\n\n\n\n\nYOU MAY BE A GOOD FIT IF YOU HAVE\n\n - You've shipped real agentic or LLM-powered systems. Not demos, but production systems with real users and real failure modes you had to debug and fix. You can talk about what broke, why, and what you changed.\n\n - You understand how retrieval, reasoning, and agent orchestration fit together as a system. Jockey's architecture sits at that intersection, and the integration surfaces you build need to reflect how that system actually behaves under real conditions.\n\n - You've owned a developer-facing or external API end-to-end, with real decisions about contracts, versioning, and the experience of building on top of your surface.\n\n - You've built authentication and authorization in production, OAuth, OIDC, RBAC, or multi-tenant isolation, as a primary owner and not a consumer of someone else's framework.\n\n - You've thought seriously about enterprise-readiness: what it takes to go from a capability that works to a platform that a large customer will trust with their data and workflows.\n\n - You bring versatility across the full platform surface. We're not looking for a specialist in one of these areas. We're looking for someone who can own all of them and make good tradeoffs across them.\n\n - You've operated in a fast-moving environment before, a startup, a hypergrowth company, or an AI-first team where you shipped under ambiguity without heavy process support.\n\n\n\n\nPREFERRED QUALIFICATIONS\n\n - Experience with MCP, A2A protocols, or the broader agent ecosystem. You've contributed to or built on these surfaces and understand where they're headed.\n\n - Hypergrowth pedigree from a technically excellent startup.\n\n\n\n\nBENEFITS AND PERKS\n\n🤝 An open and inclusive culture and work environment.\n\n🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology.\n\n🏥 Full health, dental, and vision benefits\n\n✈️ Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years.\n\n🛂 VISA support where applicable"},{"id":"08ea2b55-9c11-440f-9fbc-3195a8878fe9","title":"Staff Backend Engineer, Internal Products","department":"Product & Design","team":"Application Engineering","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-07-28T15:18:09.086+00:00","isListed":true,"isRemote":true,"workplaceType":"Hybrid","address":{"postalAddress":{"addressRegion":"CA","addressCountry":"United States","addressLocality":"San Francisco"}},"jobUrl":"https://jobs.ashbyhq.com/twelve-labs/08ea2b55-9c11-440f-9fbc-3195a8878fe9","applyUrl":"https://jobs.ashbyhq.com/twelve-labs/08ea2b55-9c11-440f-9fbc-3195a8878fe9/application","descriptionHtml":"Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
TwelveLabs builds multimodal foundation models and products for multimodal intelligence and orchestration. The quality of those models and products depends on the quality of the data and evaluation work behind them, and that work depends on the internal tools the team uses daily.
This role will own and build internal products to support model evaluation, regression testing, dataset discovery, and agentic context curation.
This is a hybrid, high autonomy role and the first on this team. You will have full ownership and drive implementation of mission-critical internal products for all members of TwelveLabs.
Unify the internal data and evaluation platform. Pull the internal beta access, evaluation, regression, and dataset discovery/cataloging products into one maintainable system that science, product, and engineering use every day.
Build a secure access layer over very large data. Make our internal data discoverable and queryable across cloud environments, at petabyte scale, with the access controls and PII handling needed to work safely with production and customer content.
Bring real customer signal into the loop. Integrate feedback and usage data from our self-serve product so the team can ground datasets and evaluations in actual customer behavior rather than abstractions.
Support the product org's internal agentic content curation. Help build and harden the internal agent the product team uses to pull context across its tools and data, and move it from prototype toward something the wider org can use, with proper access controls behind it.
Strong backend engineer with real depth in at least one language (Python, Go, or similar). You design data models, services, and APIs that hold up under real use.
Genuine system design instinct. You can take an ambiguous problem, decide what to build, and make architectural calls you can defend.
A track record of building zero to one internal products and shipping them.
Comfortable as the primary or sole engineer building the foundations for a larger team that can easily onboard onto what you’ve built.
Comfort owning a product outright: scoping it, building it, shipping it, and supporting internal users, while keeping sight of how it fits into science, product, and go to market.
Experience with large unstructured datasets.
Full stack capable. You can stand up a usable interface when the work calls for it, but your center of gravity is the backend. We pair engineers with tools like Claude Code for much of the frontend lift, so deep frontend specialization is not what this role needs.
Data platform, lakehouse, or pipeline experience across object storage, warehouses, and more than one cloud.
Experience reasoning about multimodal unstructured data (video, audio, images, text). Direct video experience is not required.
Building or integrating AI agents, MCP servers, retrieval systems, or evaluation tooling.
Data governance, access control, or PII handling on production data.
Even if there are a few checkboxes that aren’t ticked through your prior experience, we still encourage you to apply! If you are a 0-1 achiever, a ferocious learner, and a kind and fun team player who motivates others, you will find a home at TwelveLabs.
We are a global company that values the uniqueness of each person’s journey. It is the differences in our cultural, educational, and life experiences that allow us to constantly challenge the status quo. We are looking for individuals who are motivated by our mission and eager to make an impact as we push the bounds of technology to transform the world. Join us as we revolutionize video understanding and multimodal AI.
🤝 An open and inclusive culture and work environment.
🧑💻 Work closely with a collaborative, mission-driven team on cutting-edge AI technology.
🦷 Full health, dental, and vision benefits.
✈️ Flexible PTO and parental leave policy. Office closed the week of Christmas and New Years.
","descriptionPlain":"WHO WE ARE:\n\nVideo is 90% of the world's data. Most of it is invisible to machines.\nTwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.\n\n\nWe have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.\n\n\nWe are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!\n\n\n\n\nABOUT THE ROLE:\n\nTwelveLabs builds multimodal foundation models and products for multimodal intelligence and orchestration. The quality of those models and products depends on the quality of the data and evaluation work behind them, and that work depends on the internal tools the team uses daily.\n\nThis role will own and build internal products to support model evaluation, regression testing, dataset discovery, and agentic context curation.\n\nThis is a hybrid, high autonomy role and the first on this team. You will have full ownership and drive implementation of mission-critical internal products for all members of TwelveLabs.\n\n\n\n\nYOU WILL:\n\nUnify the internal data and evaluation platform. Pull the internal beta access, evaluation, regression, and dataset discovery/cataloging products into one maintainable system that science, product, and engineering use every day.\n\nBuild a secure access layer over very large data. Make our internal data discoverable and queryable across cloud environments, at petabyte scale, with the access controls and PII handling needed to work safely with production and customer content.\n\nBring real customer signal into the loop. Integrate feedback and usage data from our self-serve product so the team can ground datasets and evaluations in actual customer behavior rather than abstractions.\n\nSupport the product org's internal agentic content curation. Help build and harden the internal agent the product team uses to pull context across its tools and data, and move it from prototype toward something the wider org can use, with proper access controls behind it.\n\n\n\n\nYOU MAY BE A GOOD FIT IF YOU HAVE:\n\n - Strong backend engineer with real depth in at least one language (Python, Go, or similar). You design data models, services, and APIs that hold up under real use.\n\n - Genuine system design instinct. You can take an ambiguous problem, decide what to build, and make architectural calls you can defend.\n\n - A track record of building zero to one internal products and shipping them.\n\n - Comfortable as the primary or sole engineer building the foundations for a larger team that can easily onboard onto what you’ve built.\n\n - Comfort owning a product outright: scoping it, building it, shipping it, and supporting internal users, while keeping sight of how it fits into science, product, and go to market.\n\n - Experience with large unstructured datasets.\n\n - Full stack capable. You can stand up a usable interface when the work calls for it, but your center of gravity is the backend. We pair engineers with tools like Claude Code for much of the frontend lift, so deep frontend specialization is not what this role needs.\n\n\n\n\nYOU’LL STAND OUT IF YOU HAVE:\n\n - Data platform, lakehouse, or pipeline experience across object storage, warehouses, and more than one cloud.\n\n - Experience reasoning about multimodal unstructured data (video, audio, images, text). Direct video experience is not required.\n\n - Building or integrating AI agents, MCP servers, retrieval systems, or evaluation tooling.\n\n - Data governance, access control, or PII handling on production data.\n\n\n\nEven if there are a few checkboxes that aren’t ticked through your prior experience, we still encourage you to apply! If you are a 0-1 achiever, a ferocious learner, and a kind and fun team player who motivates others, you will find a home at TwelveLabs.\n\n \n\nWe are a global company that values the uniqueness of each person’s journey. It is the differences in our cultural, educational, and life experiences that allow us to constantly challenge the status quo. We are looking for individuals who are motivated by our mission and eager to make an impact as we push the bounds of technology to transform the world. Join us as we revolutionize video understanding and multimodal AI.\n\n \n\n\nBENEFITS AND PERKS:\n\n🤝 An open and inclusive culture and work environment.\n\n🧑💻 Work closely with a collaborative, mission-driven team on cutting-edge AI technology.\n\n🦷 Full health, dental, and vision benefits.\n\n✈️ Flexible PTO and parental leave policy. Office closed the week of Christmas and New Years.\n\n"},{"id":"be35dda7-42d3-4dbf-9b7f-3820a50727fd","title":"Senior GTM Systems Engineer","department":"Go-to-Market (GTM)","team":"Go-to-Market (GTM)","employmentType":"FullTime","location":"Remote US","secondaryLocations":[],"publishedAt":"2026-07-29T20:58:37.658+00:00","isListed":true,"isRemote":true,"workplaceType":"Remote","address":{"postalAddress":{"addressCountry":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/twelve-labs/be35dda7-42d3-4dbf-9b7f-3820a50727fd","applyUrl":"https://jobs.ashbyhq.com/twelve-labs/be35dda7-42d3-4dbf-9b7f-3820a50727fd/application","descriptionHtml":"Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
As the GTM Systems Engineer, you will own and build the systems and data foundation for TwelveLabs' go-to-market organization. You will design and ship Clay-native workflows, manage HubSpot and our marketing automation tools, build and maintain our AWS co-sell pipeline through Clazar, own quote-to-cash automation end-to-end, and build the automations that power GTM execution across sales, marketing, and partnerships. Working closely with Sales, Growth, and Partnerships leadership, you will translate GTM strategy into scalable, automated systems that operate reliably in production.
This role requires both strategic thinking and deep hands-on execution. You should be comfortable shipping workflows, managing integrations, optimizing data models, and implementing automation across multiple tools. It's ideal for someone who thrives on turning messy GTM signals into clean, automated systems and building leverage through automation.
This is a rare opportunity to be a foundational member of the TwelveLabs GTM organization. You will build the internal systems and automations that power revenue execution, working closely with GTM leadership to define how scalable, AI-native GTM infrastructure is built from the ground up.
Candidates must be able to travel up to 10% of the time annually to attend conferences, off-site meetings, and other business-related events as required by the role. This role may require participation in on-site interviews and/or completion of in-person onboarding processes.
Design, build, and ship HubSpot-based GTM systems, managing configuration, data structure, and workflow automation across our growing stack. Our current GTM stack includes HubSpot, Clay (as a core orchestration layer), Avoma, Crossbeam, Clazar, Mixpanel, and Metabase—with legacy tools like Salesloft being phased out and new tools under evaluation as we expand our AI-native GTM infrastructure.
Build and own Clay-native automations that enrich, score, route, and orchestrate GTM data in real time
Own and maintain our AWS co-sell pipeline integration (Clazar), ensuring partner-sourced opportunities sync cleanly into HubSpot
Design automation that replaces manual GTM work, including routing, lifecycle management, ownership logic, enrichment, partner attribution, and deal registration
Build reliable multi-system workflows across HubSpot, Clay, Salesloft, Avoma, Crossbeam, and Clazar that behave correctly in production
Create systems where accurate reporting is a byproduct of clean workflows and data flows, not manual effort
Prototype quickly, ship v1 systems, and iterate based on real GTM performance and adoption
Evaluate and implement new GTM tools and integrations as we expand our AI-native revenue operations infrastructure
Collaborate with Sales, Partnerships, and Growth leadership to translate go-to-market priorities into technical systems and automations
Build guardrails, QA processes, and self-healing workflows to maintain data quality and system reliability at scale
5+ years of experience building GTM systems, internal automations, or revenue infrastructure in a high-growth B2B environment
Strong technical expertise in HubSpot, including hands-on experience building workflows, data models, and integrations
Deep, practical experience with Clay beyond basic enrichment (e.g., scoring, routing, orchestration, signal detection)
Experience working with sales engagement and marketing automation tools such as HubSpot Marketing Hub, Salesloft, Outreach, or 6sense
Working knowledge of data analysis and visualization tools such as Google Sheets, Looker, Metabase, or SQL-based environments
A builder mindset — you enjoy shipping systems, automating manual work, and iterating quickly
Ability to translate complex GTM requirements into practical, scalable technical solutions
Comfort operating cross-functionally and autonomously in a fast-paced, high-growth environment
Exposure to quote-to-cash or CPQ tooling (e.g., SpotDraft) and product-led growth (PLG) motions is a plus
🤝 An open and inclusive culture and work environment
🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology
🏥 Full health, dental, and vision benefits
🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years
💪 Monthly wellness stipend
📚 Annual Learning & Development stipend to invest in your growth
💼 Global offices in San Francisco and Seoul, and coworking office memberships for remote team members
🛂 VISA support where applicable
🚆 Transportation stipend (SF onsite only)
🍲 Daily lunch & dinner provided (SF onsite only)
","descriptionPlain":"WHO WE ARE\n\nVideo is 90% of the world's data. Most of it is invisible to machines.\nTwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.\n\n\nWe have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.\n\n\nWe are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!\n\n\n\n\nABOUT THE ROLE\n\nAs the GTM Systems Engineer, you will own and build the systems and data foundation for TwelveLabs' go-to-market organization. You will design and ship Clay-native workflows, manage HubSpot and our marketing automation tools, build and maintain our AWS co-sell pipeline through Clazar, own quote-to-cash automation end-to-end, and build the automations that power GTM execution across sales, marketing, and partnerships. Working closely with Sales, Growth, and Partnerships leadership, you will translate GTM strategy into scalable, automated systems that operate reliably in production.\n\n\n\nThis role requires both strategic thinking and deep hands-on execution. You should be comfortable shipping workflows, managing integrations, optimizing data models, and implementing automation across multiple tools. It's ideal for someone who thrives on turning messy GTM signals into clean, automated systems and building leverage through automation.\n\n\nThis is a rare opportunity to be a foundational member of the TwelveLabs GTM organization. You will build the internal systems and automations that power revenue execution, working closely with GTM leadership to define how scalable, AI-native GTM infrastructure is built from the ground up.\n\n\n\nCandidates must be able to travel up to 10% of the time annually to attend conferences, off-site meetings, and other business-related events as required by the role. This role may require participation in on-site interviews and/or completion of in-person onboarding processes.\n\n\n\n\nIN THIS ROLE, YOU WILL\n\n - Design, build, and ship HubSpot-based GTM systems, managing configuration, data structure, and workflow automation across our growing stack. Our current GTM stack includes HubSpot, Clay (as a core orchestration layer), Avoma, Crossbeam, Clazar, Mixpanel, and Metabase—with legacy tools like Salesloft being phased out and new tools under evaluation as we expand our AI-native GTM infrastructure.\n\n - Build and own Clay-native automations that enrich, score, route, and orchestrate GTM data in real time\n\n - Own and maintain our AWS co-sell pipeline integration (Clazar), ensuring partner-sourced opportunities sync cleanly into HubSpot\n\n - Design automation that replaces manual GTM work, including routing, lifecycle management, ownership logic, enrichment, partner attribution, and deal registration\n\n - Build reliable multi-system workflows across HubSpot, Clay, Salesloft, Avoma, Crossbeam, and Clazar that behave correctly in production\n\n - Create systems where accurate reporting is a byproduct of clean workflows and data flows, not manual effort\n\n - Prototype quickly, ship v1 systems, and iterate based on real GTM performance and adoption\n\n - Evaluate and implement new GTM tools and integrations as we expand our AI-native revenue operations infrastructure\n\n - Collaborate with Sales, Partnerships, and Growth leadership to translate go-to-market priorities into technical systems and automations\n\n - Build guardrails, QA processes, and self-healing workflows to maintain data quality and system reliability at scale\n \n\n\nYOU MAY BE A GOOD FIT IF YOU HAVE\n\n - 5+ years of experience building GTM systems, internal automations, or revenue infrastructure in a high-growth B2B environment\n\n - Strong technical expertise in HubSpot, including hands-on experience building workflows, data models, and integrations\n\n - Deep, practical experience with Clay beyond basic enrichment (e.g., scoring, routing, orchestration, signal detection)\n\n - Experience working with sales engagement and marketing automation tools such as HubSpot Marketing Hub, Salesloft, Outreach, or 6sense\n\n - Working knowledge of data analysis and visualization tools such as Google Sheets, Looker, Metabase, or SQL-based environments\n\n - A builder mindset — you enjoy shipping systems, automating manual work, and iterating quickly\n\n - Ability to translate complex GTM requirements into practical, scalable technical solutions\n\n - Comfort operating cross-functionally and autonomously in a fast-paced, high-growth environment\n\n - Exposure to quote-to-cash or CPQ tooling (e.g., SpotDraft) and product-led growth (PLG) motions is a plus\n\n\n\n\nBENEFITS AND PERKS\n\n🤝 An open and inclusive culture and work environment\n\n🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology\n\n🏥 Full health, dental, and vision benefits\n\n🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years\n\n💪 Monthly wellness stipend\n\n📚 Annual Learning & Development stipend to invest in your growth\n\n💼 Global offices in San Francisco and Seoul, and coworking office memberships for remote team members\n\n🛂 VISA support where applicable\n\n🚆 Transportation stipend (SF onsite only)\n\n🍲 Daily lunch & dinner provided (SF onsite only)"},{"id":"cb6f2b66-e266-4534-bd8d-3619408c91fc","title":"Tech Lead Manager, Jockey Core","department":"Tech","team":"ML Engineering","employmentType":"FullTime","location":"Seoul, South Korea","secondaryLocations":[],"publishedAt":"2026-08-04T13:56:43.892+00:00","isListed":true,"isRemote":true,"workplaceType":"Hybrid","address":{"postalAddress":{"addressCountry":"South Korea","addressLocality":"Seoul"}},"jobUrl":"https://jobs.ashbyhq.com/twelve-labs/cb6f2b66-e266-4534-bd8d-3619408c91fc","applyUrl":"https://jobs.ashbyhq.com/twelve-labs/cb6f2b66-e266-4534-bd8d-3619408c91fc/application","descriptionHtml":"Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.
No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.
Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.
We build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.
Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.
The Cognition Models team owns the models that turn video into structured understanding and reasoning: Pegasus, our video-language model, and Jockey Core, the reasoning LLM behind Jockey. In the model stack we sit between Perception Models (embeddings and retrieval) and the agent system — taking what's retrieved and producing structured understanding and the reasoning to act on it.
We focus on multimodal systems with high instruction-following capability and complex, hierarchically structured outputs. Our work spans training infrastructure from pre-training to RL, temporal segmentation and structured metadata extraction, large-scale inference and serving systems, data-curation and evaluation pipelines, and building Jockey Core. We ship products with real-world value rather than doing research in isolation, working as a goal-oriented, cross-functional team of ML researchers and engineers — using the most advanced compute in the world, including NVIDIA B300s, to accelerate the research-to-production cycle.
Jockey Core is the reasoning LLM at the center of Jockey — the model that decomposes a query, decides what to retrieve and segment, and reasons over the results into an answer you can act on. It sits in the critical path of every agent step, so its quality, latency, and cost directly shape what Jockey can do. Jockey Core is a model we own and serve end to end, and we improve it continuously so Jockey's quality compounds with every release.
This is a Tech Lead Manager role to build and lead a newly founded team building Jockey Core — a leader who stays deeply hands-on while standing up and growing the team.
Build and lead the founding team — hiring, growth, delivery, and technical direction.
Own Jockey Core's end-to-end roadmap, from model and engine selection through model-efficiency work (pruning, quantization, distillation) to production serving and scale-out.
Stay hands-on: lead critical system and serving/inference architecture decisions, and set the technical bar through design review.
Judge latency/throughput/cost tradeoffs with measured data, and partner with the Pegasus, agent, and infrastructure teams on capacity and SLOs.
Explore and adopt AI-assisted development tools (Claude, Gemini, GPT) to raise team productivity.
A track record leading ML/infrastructure teams as a hands-on tech lead or manager — ideally founding or scaling a team from small.
Deep experience serving and optimizing large-scale LLM inference in production (vLLM, TensorRT-LLM, SGLang, or similar), across techniques like batching/scheduling, quantization, disaggregated prefill/decode, and speculative decoding.
A habit of driving ambiguous technical decisions with measured latency/throughput/cost data.
Excellent communication and people leadership.
Experience with model compression (pruning, quantization-aware training, distillation) or productionizing reasoning/agentic LLMs.
Experience with multi-region/multi-cluster serving or large-scale GPU capacity planning.
Contributions to or customization of an LLM inference server's internals.
A Master's/PhD in Machine Learning, Computer Science, or a related field.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
영상은 전 세계 데이터의 90%를 차지하지만, 그 대부분은 아직 기계가 이해할 수 없는 영역에 머물러 있습니다.
트웰브랩스(TwelveLabs)는 이 문제를 해결하기 위해 기계가 영상을 이해할 수 있도록 하는 AI 인프라를 구축합니다. 사람보다 더 정교하고 깊이 있게 영상을 이해하는 AI 모델을 통해 영상 속 시·청각 정보를 종합 분석하여 미디어·엔터테인먼트, 스포츠, 보안, 공공 분야 전반의 워크플로우에 적용하고 있습니다.
NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, 네이버벤처스, 한국투자파트너스, Quadrille Capital, Red Bull Ventures 등 글로벌 투자자들로부터 누적 3천억원 이상의 투자를 유치했습니다. 또한 Fei-Fei Li, Silvio Savarese, Alexandr Wang 등 세계적인 AI 리더들이 자문진으로 함께하고 있습니다.
트웰브랩스는 샌프란시스코, 서울, 뉴욕, 런던 네 개의 오피스를 거점으로 운영되는 글로벌 기업입니다. 핵심 연구개발은 서울에서 이루어지며, 전 세계 오피스의 구성원들이 이를 기반으로 다양한 산업과 시장에 영상 이해 기술을 적용해가고 있습니다. 세상의 복잡함을 이해하는 기술을 만드는 만큼, 서로 다른 배경과 관점을 가진 사람들이 모일 때 더 나은 제품을 만들 수 있다고 믿습니다.
아직 세상에 없는 답을 정의해보고 싶으신 분, 내 손으로 직접 변화를 만들어내고 싶어하시는 분을 찾고 있습니다. 트웰브랩스에서 세상을 이해하는 기술을 함께 만들어가세요.
토큰(token)은 대규모 언어 모델과 AI 코딩 도구가 입력과 출력을 처리하는 사용량 단위입니다. Token Engineer는 AI 사용을 단순히 줄이거나 비용을 사후 점검하는 역할이 아닙니다. 연구자와 엔지니어가 AI를 적극적으로 활용하면서도 비용, 성능, 보안, 안정성을 일관되게 관리할 수 있도록 토큰 관리 및 최적화 솔루션을 만드는 역할입니다.
여러 AI 도구와 모델 제공자가 각기 다른 사용량 및 비용 정보를 제공하는 환경에서, 이 역할은 요청 단위의 계측부터 중앙 정책과 자동 통제까지 이어지는 하나의 시스템을 설계하고 구현합니다. 어떤 도구와 모델이 실제 업무 성과에 도움이 되는지 측정하고, 비정상적인 사용을 빠르게 감지하며, 근거 있는 사용 기준을 운영 가능한 정책으로 바꿉니다.
초기 설계부터 프로덕션 운영까지 직접 책임지게 됩니다. Infrastructure, Security, Finance, IT, Research, Product, Engineering과 협업하면서, 사람의 수동적인 개입에 의존하지 않는 안전하고 투명한 AI 사용 환경을 구축합니다.
여러 AI 코딩 도구와 모델 API의 토큰 사용량, 비용, 모델, 지연 시간, 오류, 재시도, 사용자 및 팀 정보를 하나의 표준 데이터 모델로 수집하는 계측 파이프라인을 설계하고 구현합니다.
일별 사용량과 비용을 거의 실시간으로 파악할 수 있는 대시보드, 자동 알림, 이상 징후 탐지 시스템을 구축하고, 제공자 청구 내역과 내부 계측 결과가 일치하는지 지속적으로 검증합니다.
인증, 권한, 사용 한도, 예산 기준, 요청 속도 제한, 모델 및 제공자 라우팅, 캐시, 재시도, 감사 로그를 중앙에서 관리하는 AI 게이트웨이와 제어 계층을 설계하고 운영합니다.
과거 사용 데이터와 실제 업무 효과를 바탕으로 합리적인 기준선을 만들고, 생산성과 실험 속도를 해치지 않으면서 과도한 비용과 운영 위험을 막는 정책을 정의합니다.
원시 토큰 수나 청구 금액만 보지 않고, 성공한 작업당 비용, 지연 시간, 실패율, 재시도율처럼 실제 효율을 설명하는 지표를 설계합니다.
AWS와 Kubernetes 환경에서 안전하고 확장 가능한 서비스를 구축하고, Infrastructure as Code, 비밀정보 관리, 역할 기반 접근 제어, 데이터 보존 정책을 적용합니다.
시스템의 서비스 수준 목표, 운영 절차, 장애 대응 체계를 정의하고, 비용 급증이나 제공자 장애가 발생했을 때 원인을 빠르게 좁히고 복구할 수 있는 도구를 만듭니다.
각 팀이 자신의 사용 현황과 개선 기회를 스스로 이해할 수 있는 셀프서비스 도구와 문서를 제공하고, 반복적인 수동 확인을 자동화합니다.
새로운 모델과 제공자의 품질, 비용, 지연 시간, 안정성을 비교하고, 업무 특성에 맞는 라우팅과 최적화 방안을 제안하고 구현합니다.
다양한 직군의 요구를 기술 설계로 바꾸고, 우선순위와 상충 관계를 명확히 설명하며, 모호한 문제를 실제 운영 가능한 시스템으로 끝까지 완성합니다.
인프라, 플랫폼, 백엔드 또는 엔지니어링 분야에서 5년 이상 일했거나, 이에 준하는 깊이로 프러덕션 환경을 구축하고 운영한 경험이 있으신 분
AWS와 Kubernetes를 활용해 실제 서비스를 구축하고 운영했으며, Terraform과 같은 Infrastructure as Code 도구로 반복 가능한 환경을 만든 경험이 있으신 분
Python 또는 Go로 운영 도구가 아닌 장기적으로 유지되는 프로덕션 서비스를 개발할 수 있으신 분
로그, 메트릭, 분산 추적과 같은 관측성 데이터를 설계하고, 장애나 성능 저하의 근본 원인을 데이터로 찾아낸 경험이 있으신 분
API 게이트웨이, 프록시, 네트워크, 인증 및 권한 관리 중 하나 이상의 영역을 깊이 다뤄본 경험이 있으신 분
사용량 계측, 과금, 할당량, 비용 배분 또는 대규모 이벤트 처리 시스템을 설계하거나 운영한 경험이 있으신 분
보안, 안정성, 비용, 개발자 경험 사이의 상충 관계를 이해하고, 현실적인 설계 결정을 내릴 수 있으신 분
요구사항이 완전히 정리되지 않은 상황에서도 문제를 구조화하고, 설계부터 배포와 운영까지 주도적으로 완성한 경험이 있으신 분
기술적 내용을 엔지니어와 비전문가 모두에게 명확하게 설명하고, 여러 조직과 합의점을 만들어낼 수 있으신 분
단기적인 수동 대응에 머무르지 않고, 같은 문제가 반복되지 않도록 자동화와 시스템 개선으로 연결하는 분
대규모 언어 모델 API, AI 게이트웨이, 모델 서빙 또는 AI 코딩 도구를 위한 내부 플랫폼을 구축한 경험
OpenTelemetry, Prometheus, Grafana, Loki, Tempo 등으로 관측성 플랫폼을 구축하고 운영한 경험
클라우드와 AI 비용을 기술적으로 측정하고 최적화하는 FinOps, 비용 귀속, 내부 정산 체계를 다뤄본 경험
요청별 사용량 계측, 구독 및 과금, 할당량, 속도 제한, 정책 엔진을 구현한 경험
OIDC, SSO, SCIM, 역할 기반 접근 제어, 비밀정보 관리 및 감사 로그를 포함한 엔터프라이즈 보안 경험
시계열 데이터의 이상 징후 탐지, 비용 예측 또는 용량 계획을 실제 운영에 적용한 경험
여러 모델 제공자를 대상으로 품질, 비용, 지연 시간에 따라 요청을 라우팅하거나 장애 시 우회 경로를 설계한 경험
개발자 플랫폼이나 셀프서비스 인프라를 만들어 조직 전체의 운영 부담을 줄인 경험
명시된 모든 요건을 충족하지 않더라도, 복잡한 인프라 문제를 깊이 파고들어 실제로 운영되는 시스템으로 완성해 온 분이라면 지원을 권합니다.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
With our Series B, we are entering our next chapter: building a full-stack video AI company with frontier models at the foundation, an enterprise platform in the middle, and agentic systems that turn raw video into actionable knowledge and insights.
Our customers include some of the world’s leading companies across media, sports, advertising, security, and the public sector. They are asking us to solve increasingly complex problems—not only through our APIs, but by deploying TwelveLabs technology securely and reliably into their environments.
As Director of Engineering, Deployment & Integrations, you will lead the engineering organization responsible for deploying TwelveLabs technology into our customers' and partners’ environments and integrating it deeply into their existing technology ecosystems.
You will own the engineering strategy and execution for customer and distribution partner deployments across hyperscalers, private cloud, on-premise, and other complex enterprise environments. You will work closely with our Engineering and Go to Market teams to turn complex customer requirements into scalable, repeatable deployment and integration capabilities.
This is a highly cross-functional and customer-facing engineering leadership role. You will be equally comfortable working through architecture with a customer's technical team, diving into a complex deployment or integration challenge, and building the systems and processes that allow the next customer to move faster.
Your goal is to make TwelveLabs easy to deploy, easy to integrate, and reliable to operate — while turning the lessons from each customer into capabilities that scale across the business.
Customer Deployment
Own the engineering lifecycle for customer and partner deployments, from technical discovery and architecture through implementation, production launch, and ongoing operations.
Lead deployments across public cloud, customer VPCs, private cloud, on-premise infrastructure, and other customer-controlled environments.
Partner directly with strategic customers, partners and their engineering, infrastructure, security, and IT teams to solve complex deployment challenges.
Establish the technical standards, tooling, and playbooks required to make deployments predictable and repeatable.
Turn recurring customer deployment requirements into scalable product and platform capabilities wherever possible.
Integrations & Enterprise Architecture
Own the engineering strategy for integrating TwelveLabs into customers' existing technology ecosystems.
Build scalable integration patterns across enterprise applications, data platforms, storage systems, identity and access management, APIs, and AI/ML infrastructure.
Partner with Product and Engineering to translate recurring customer integration requirements into reusable platform capabilities.
Establish standards and tooling that make integrations faster, more reliable, and increasingly self-service.
Balance customer-specific requirements with a long-term architecture that prevents TwelveLabs from accumulating one-off integrations.
Help define the technical architecture and interfaces that allow TwelveLabs to become deeply embedded in customer workflows and applications.
Deployment Automation & Developer Experience
Build the tooling, automation, and internal platforms that allow engineering teams to provision, configure, upgrade, integrate, and operate customer environments efficiently.
Develop deployment and integration architectures that move TwelveLabs from bespoke implementations toward repeatable, highly automated capabilities.
Establish strong CI/CD, release management, configuration management, and observability practices for distributed customer environments.
Create an exceptional internal developer experience for engineers responsible for building and supporting deployments and integrations.
Product & Engineering Partnership
Work closely with Product and Engineering leadership to ensure deployment and integration requirements are incorporated into product architecture from the beginning.
Identify recurring friction across customer deployments and integrations and turn those insights into product and platform improvements.
Partner with Research and ML teams to understand the operational requirements of serving increasingly capable multimodal models in production environments.
Help define the technical roadmap for deployment and integration infrastructure as TwelveLabs expands its product portfolio and customer footprint.
Customer, Partnerships & Cross-Functional Leadership
Serve as a senior technical partner to Solutions Engineering, Sales, and Product on the most complex customer opportunities.
Translate customer requirements into clear technical decisions, engineering priorities, and execution plans.
Establish a strong operating rhythm across customer-facing and engineering teams to ensure deployments and integrations move quickly without compromising quality.
Communicate complex technical tradeoffs clearly to both highly technical audiences and executive stakeholders.
Security, Compliance & Trust
Build deployment and integration architectures suitable for enterprise and public sector customers with demanding security and compliance requirements.
Partner with Security and Infrastructure teams to support requirements around data isolation, identity, access control, encryption, auditability, and compliance.
Help establish deployment patterns capable of supporting increasingly regulated and security-sensitive environments.
Team Building & Culture
Build and lead a world-class Deployment & Integrations Engineering organization.
Recruit, develop, and retain engineers who thrive at the intersection of distributed systems, infrastructure, product engineering, enterprise integrations, and customer problem-solving.
Create a culture of high ownership, technical excellence, urgency, and disciplined execution.
Establish the processes and operating mechanisms that allow the team to move quickly while maintaining a high bar for reliability and customer experience.
10+ years of experience in software engineering, infrastructure, platform engineering, or distributed systems, with significant experience leading globally distributed engineering teams.
Experience leading engineering organizations responsible for deploying complex software, AI/ML systems, or infrastructure into enterprise customer environments.
Deep technical understanding of cloud infrastructure, distributed systems, networking, containers, Kubernetes, CI/CD, observability, and production operations.
Experience with Kubernetes, container orchestration, infrastructure-as-code, and automated deployment systems.
Experience deploying software across multiple cloud environments and/or customer-controlled infrastructure, including air-gapped, disconnected, regulated, or highly secure environments
Hands-on experience owning the software deployment lifecycle end-to-end — versioning, release management, rollout + rollback strategy, upgrade paths, and post-deployment operations across distributed customer environments.
Deep experience with software packaging patterns for shipping into customer-controlled environments — container images, Helm charts, operator patterns, installer + bootstrap flows, air-gapped bundles, and dependency + configuration management.
A track record of turning bespoke customer implementations into scalable, repeatable engineering systems.
Strong customer-facing instincts and the ability to operate credibly with senior engineers, architects, security teams, and technical executives at large enterprises.
A bias toward action and a willingness to get into the technical details when the situation demands it.
Excellent judgment about when to build a one-off solution for a strategic customer versus investing in a scalable platform capability.
Experience operating in a high-growth startup where product requirements and customer needs evolve rapidly.
Experience working with globally distributed teams
Experience deploying AI/ML or GPU-intensive workloads into customer environments.
Experience with public sector, defense, financial services, or other highly regulated customers.
Experience taking an enterprise deployment motion from bespoke implementations to a standardized platform.
Experience building or leading Solutions Engineering, Field Engineering, Deployment Engineering, or Professional Services Engineering organizations.
We are a global company that values the uniqueness of each person’s journey. The differences in our cultural, educational, and life experiences are what allow us to challenge the status quo. We’re looking for people motivated by our mission and ready to help build the company that defines how the world understands video.
🤝 An open and inclusive culture and work environment
🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology
🏥 Full health, dental, and vision benefits
🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years
💪 Monthly wellness stipend
📚 Annual Learning & Development stipend to invest in your growth
💼 Global offices in San Francisco, New York, and Seoul, and coworking office memberships for remote team members
🛂 VISA support where applicable
🚆 Transportation stipend (SF & NY onsite only)
🍲 Daily meals provided (SF & NY onsite only)
","descriptionPlain":"WHO WE ARE\n\n\n\nVideo is 90% of the world's data. Most of it is invisible to machines.\nTwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.\n\n\nWe have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.\n\n\nWe are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!\n\n\n\n\nABOUT THE ROLE\n\n\n\nWith our Series B, we are entering our next chapter: building a full-stack video AI company with frontier models at the foundation, an enterprise platform in the middle, and agentic systems that turn raw video into actionable knowledge and insights.\n\n\n\nOur customers include some of the world’s leading companies across media, sports, advertising, security, and the public sector. They are asking us to solve increasingly complex problems—not only through our APIs, but by deploying TwelveLabs technology securely and reliably into their environments.\n\n\n\nAs Director of Engineering, Deployment & Integrations, you will lead the engineering organization responsible for deploying TwelveLabs technology into our customers' and partners’ environments and integrating it deeply into their existing technology ecosystems.\n\n\n\nYou will own the engineering strategy and execution for customer and distribution partner deployments across hyperscalers, private cloud, on-premise, and other complex enterprise environments. You will work closely with our Engineering and Go to Market teams to turn complex customer requirements into scalable, repeatable deployment and integration capabilities.\n\n\n\nThis is a highly cross-functional and customer-facing engineering leadership role. You will be equally comfortable working through architecture with a customer's technical team, diving into a complex deployment or integration challenge, and building the systems and processes that allow the next customer to move faster.\n\n\n\nYour goal is to make TwelveLabs easy to deploy, easy to integrate, and reliable to operate — while turning the lessons from each customer into capabilities that scale across the business.\n\n\n\n\nIN THIS ROLE, YOU WILL FOCUS ON THE CORE COMPETENCIES LISTED BELOW\n\n\n\n\nCustomer Deployment\n\n - Own the engineering lifecycle for customer and partner deployments, from technical discovery and architecture through implementation, production launch, and ongoing operations.\n\n - Lead deployments across public cloud, customer VPCs, private cloud, on-premise infrastructure, and other customer-controlled environments.\n\n - Partner directly with strategic customers, partners and their engineering, infrastructure, security, and IT teams to solve complex deployment challenges.\n\n - Establish the technical standards, tooling, and playbooks required to make deployments predictable and repeatable.\n\n - Turn recurring customer deployment requirements into scalable product and platform capabilities wherever possible.\n\nIntegrations & Enterprise Architecture\n\n - Own the engineering strategy for integrating TwelveLabs into customers' existing technology ecosystems.\n\n - Build scalable integration patterns across enterprise applications, data platforms, storage systems, identity and access management, APIs, and AI/ML infrastructure.\n\n - Partner with Product and Engineering to translate recurring customer integration requirements into reusable platform capabilities.\n\n - Establish standards and tooling that make integrations faster, more reliable, and increasingly self-service.\n\n - Balance customer-specific requirements with a long-term architecture that prevents TwelveLabs from accumulating one-off integrations.\n\n - Help define the technical architecture and interfaces that allow TwelveLabs to become deeply embedded in customer workflows and applications.\n\nDeployment Automation & Developer Experience\n\n - Build the tooling, automation, and internal platforms that allow engineering teams to provision, configure, upgrade, integrate, and operate customer environments efficiently.\n\n - Develop deployment and integration architectures that move TwelveLabs from bespoke implementations toward repeatable, highly automated capabilities.\n\n - Establish strong CI/CD, release management, configuration management, and observability practices for distributed customer environments.\n\n - Create an exceptional internal developer experience for engineers responsible for building and supporting deployments and integrations.\n\nProduct & Engineering Partnership\n\n - Work closely with Product and Engineering leadership to ensure deployment and integration requirements are incorporated into product architecture from the beginning.\n\n - Identify recurring friction across customer deployments and integrations and turn those insights into product and platform improvements.\n\n - Partner with Research and ML teams to understand the operational requirements of serving increasingly capable multimodal models in production environments.\n\n - Help define the technical roadmap for deployment and integration infrastructure as TwelveLabs expands its product portfolio and customer footprint.\n\nCustomer, Partnerships & Cross-Functional Leadership\n\n - Serve as a senior technical partner to Solutions Engineering, Sales, and Product on the most complex customer opportunities.\n\n - Translate customer requirements into clear technical decisions, engineering priorities, and execution plans.\n\n - Establish a strong operating rhythm across customer-facing and engineering teams to ensure deployments and integrations move quickly without compromising quality.\n\n - Communicate complex technical tradeoffs clearly to both highly technical audiences and executive stakeholders.\n\nSecurity, Compliance & Trust\n\n - Build deployment and integration architectures suitable for enterprise and public sector customers with demanding security and compliance requirements.\n\n - Partner with Security and Infrastructure teams to support requirements around data isolation, identity, access control, encryption, auditability, and compliance.\n\n - Help establish deployment patterns capable of supporting increasingly regulated and security-sensitive environments.\n\nTeam Building & Culture\n\n - Build and lead a world-class Deployment & Integrations Engineering organization.\n\n - Recruit, develop, and retain engineers who thrive at the intersection of distributed systems, infrastructure, product engineering, enterprise integrations, and customer problem-solving.\n\n - Create a culture of high ownership, technical excellence, urgency, and disciplined execution.\n\n - Establish the processes and operating mechanisms that allow the team to move quickly while maintaining a high bar for reliability and customer experience.\n\n\n\n\nYOU MAY BE A GOOD FIT IF YOU HAVE\n\n - 10+ years of experience in software engineering, infrastructure, platform engineering, or distributed systems, with significant experience leading globally distributed engineering teams.\n\n - Experience leading engineering organizations responsible for deploying complex software, AI/ML systems, or infrastructure into enterprise customer environments.\n\n - Deep technical understanding of cloud infrastructure, distributed systems, networking, containers, Kubernetes, CI/CD, observability, and production operations.\n\n - Experience with Kubernetes, container orchestration, infrastructure-as-code, and automated deployment systems.\n\n - Experience deploying software across multiple cloud environments and/or customer-controlled infrastructure, including air-gapped, disconnected, regulated, or highly secure environments\n\n - Hands-on experience owning the software deployment lifecycle end-to-end — versioning, release management, rollout + rollback strategy, upgrade paths, and post-deployment operations across distributed customer environments.\n\n - Deep experience with software packaging patterns for shipping into customer-controlled environments — container images, Helm charts, operator patterns, installer + bootstrap flows, air-gapped bundles, and dependency + configuration management.\n\n - A track record of turning bespoke customer implementations into scalable, repeatable engineering systems.\n\n - Strong customer-facing instincts and the ability to operate credibly with senior engineers, architects, security teams, and technical executives at large enterprises.\n\n - A bias toward action and a willingness to get into the technical details when the situation demands it.\n\n - Excellent judgment about when to build a one-off solution for a strategic customer versus investing in a scalable platform capability.\n\n - Experience operating in a high-growth startup where product requirements and customer needs evolve rapidly.\n\n - Experience working with globally distributed teams\n\n\n\n\nSTRONG CANDIDATES MAY ALSO HAVE\n\n - Experience deploying AI/ML or GPU-intensive workloads into customer environments.\n\n - Experience with public sector, defense, financial services, or other highly regulated customers.\n\n - Experience taking an enterprise deployment motion from bespoke implementations to a standardized platform.\n\n - Experience building or leading Solutions Engineering, Field Engineering, Deployment Engineering, or Professional Services Engineering organizations.\n \n \n\nWe are a global company that values the uniqueness of each person’s journey. The differences in our cultural, educational, and life experiences are what allow us to challenge the status quo. We’re looking for people motivated by our mission and ready to help build the company that defines how the world understands video.\n\n\n\n\nBENEFITS AND PERKS\n\n🤝 An open and inclusive culture and work environment\n\n🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology\n\n🏥 Full health, dental, and vision benefits\n\n🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years\n\n💪 Monthly wellness stipend\n\n📚 Annual Learning & Development stipend to invest in your growth\n\n💼 Global offices in San Francisco, New York, and Seoul, and coworking office memberships for remote team members\n\n🛂 VISA support where applicable\n\n🚆 Transportation stipend (SF & NY onsite only)\n\n🍲 Daily meals provided (SF & NY onsite only)"},{"id":"a8fa59d5-e364-43ec-9a5a-dc98a2f3ed53","title":"Senior Machine Learning Engineer, Pegasus","department":"Tech","team":"ML Engineering","employmentType":"FullTime","location":"Seoul, South Korea","secondaryLocations":[],"publishedAt":"2026-08-19T02:12:53.604+00:00","isListed":true,"isRemote":true,"workplaceType":"Hybrid","address":{"postalAddress":{"addressCountry":"South Korea","addressLocality":"Seoul"}},"jobUrl":"https://jobs.ashbyhq.com/twelve-labs/a8fa59d5-e364-43ec-9a5a-dc98a2f3ed53","applyUrl":"https://jobs.ashbyhq.com/twelve-labs/a8fa59d5-e364-43ec-9a5a-dc98a2f3ed53/application","descriptionHtml":"Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.
No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.
Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.
We build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.
Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.
The Cognition Models team owns the models that turn video into structured understanding and reasoning: Pegasus, our video-language model, and Jockey Core, the reasoning LLM behind Jockey. In the model stack we sit between Perception Models (embeddings and retrieval) and the agent system — taking what's retrieved and producing structured understanding and the reasoning to act on it.
We focus on multimodal systems with high instruction-following capability and complex, hierarchically structured outputs. Our work spans training infrastructure from pre-training to RL, temporal segmentation and structured metadata extraction, large-scale inference and serving systems, data-curation and evaluation pipelines, and building Jockey Core. We ship products with real-world value rather than doing research in isolation, working as a goal-oriented, cross-functional team of ML researchers and engineers — using the most advanced compute in the world, including NVIDIA B300s, to accelerate the research-to-production cycle.
Pegasus is TwelveLabs' video-language model — it turns video into useful analysis by reasoning over visuals, speech, audio, and on-screen text. A key capability is Segment, our time-based metadata feature: instead of a broad question about a video, customers define the exact segment types they care about and the metadata fields they want back, and Pegasus finds the relevant start and end times and returns structured metadata for each segment — titles, summaries, topics, people, visual subjects, confidence, or domain-specific labels. This turns video into time-based, structured data that flows directly into search, archive, editing, compliance, or content-management workflows.
Build, improve, and operate production ML systems for Pegasus, with a focus on reliability, performance, and maintainability.
Work across core parts of the ML stack, including deployment, inference, evaluation, monitoring, and supporting infrastructure.
Develop systems for serving Video Language Models (VLMs) and handling multimodal data and metadata at production quality.
Make strong technical decisions within your area and drive execution with a high degree of ownership.
Explore and adopt AI-assisted development tools such as Claude, Gemini, and GPT to improve productivity across coding, experimentation, debugging, and documentation.
Strong software engineering and machine learning fundamentals.
Experience building and shipping ML systems in production.
Experience with multimodal data and familiarity with areas such as computer vision, natural language processing, LLMs, or VLMs.
Experience with distributed ML or data workflows, ideally in Kubernetes-based environments.
Strong engineering judgment around performance, reliability, and maintainability in production environments.
Experience serving or optimizing LLM/VLM systems in production.
Experience with inference optimization techniques such as batching, caching, or quantization.
Experience building AI/ML systems from early-stage development through production deployment.
Master's or PhD in Machine Learning, Computer Science, or a related technical field.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
WHO WE ARE
Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised over $200M from Silicon Valley's best VCs, including Index Ventures and NEA, and from leading enterprises like Nvidia and Amazon that are shaping the future of AI.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
ABOUT THE ROLE
TwelveLabs spent years building the intelligence layer for video: shared infrastructure, multimodal video models, and an agent harness — the orchestration layer — that can understand and reason over video. Our vision for video superintelligence and pioneering work for large-scale video understanding allowed us to raise over $200M from Silicon Valley's best VCs like Index Ventures and NEA and leading enterprises like Nvidia and Amazon that are shaping the future of AI. The next chapter for TwelveLabs is to turn that foundation into products that people can use without needing to understand the technology underneath.
We are hiring a Head of Applications to build that layer.
You will decide where TwelveLabs should play, find the first use cases where our technology creates a step change in the user experience, and turn those insights into products customers use and pay for. You will own the application layer from customer discovery through product, engineering, launch, adoption, and growth.
This is a builder-GM role. It is not a conventional product management job, a growth marketing job, or a large-team executive role. You should be comfortable talking with customers in the morning, working through a product or technical tradeoff in the afternoon, and assembling an exceptional team for the 0-to-1 journey that evening. Ideal candidates will come with 12+ years of experience across product and product leadership roles, preferably at early to mid stage software companies.
IN THIS ROLE, YOU WILL
Set the strategy for TwelveLabs' application layer: which users and workflows to serve, what to build, and what not to build.
Find a focused wedge through direct customer discovery, rapid prototyping, and disciplined market testing.
Ship opinionated, AI-native applications for non-technical users on top of TwelveLabs' core models, agent harness, and shared infrastructure. Define application-specific evals, task-success criteria, reliability thresholds, human review, and launch guardrails.
Own the full outcome of the applications you ship, including activation, engagement, retention, monetization, and customer value. Shipping is the start, not the finish.
Build a fast learning loop between users and the product. Turn usage data and customer behavior into better workflows, interfaces, and model experiences.
Act as the demanding internal customer for our platform and model teams. Define the capabilities, APIs, latency, quality, and reliability the application layer needs without rebuilding the layers below it.
Partner closely with Design, Research and Engineering, Product, Commercial, Revenue, and Partnerships to take products from prototype to repeatable adoption. Work deeply with lighthouse customers, while converting what you learn into reusable products rather than one-off implementations.
Recruit and lead a lean and extremely high-performing team with the range to move from idea to production.
Establish a culture of speed, craft, direct user contact, and clear accountability. Keep the team lean as the product and business grow.
You May be a good fit if you have
You have built and grown a product from zero to meaningful adoption. You can explain what users did, what changed because of your decisions, and where you were wrong.
You have exceptional product judgment and know what great looks like. You can find the narrow workflow that matters, reduce a complex technology to a simple experience, and say no to attractive distractions.
You are technically fluent enough to work directly with engineers and researchers on architecture, model behavior, latency, evaluation, reliability, and cost. You do not need to be the strongest engineer in the room, but you cannot treat the technology as a black box.
You understand the realities of AI-native products: probabilistic behavior, evaluation design, trust, human review, privacy and rights constraints, and the tradeoffs between quality, latency, and cost.
You understand growth as part of the product. You have designed or led loops that improved activation, retention, distribution, or monetization, not just acquisition campaigns.
You recruit well and have built an early team of unusually strong, low-ego people. You know how to operate before every function has its own leader.
You are close to the work. You use the product, talk with users, inspect the data, and raise the quality bar through direct involvement.
You move quickly without confusing activity with progress. You create evidence, make decisions, and change course when the evidence says you should.
You communicate clearly across research, engineering, design, and go-to-market teams and can resolve cross-layer tradeoffs without creating organizational drag.
WHAT SUCCESS LOOKS LIKE IN THE FIRST 90 DAYS
Build a clear view of the most promising users, workflows, and distribution paths for the applications built.
Narrow the field to one or two application wedges using an explicit market-attractiveness and right-to-win filter, with clear reasons to reject the rest.
Put working prototypes in users' hands and establish a weekly build-measure-learn cadence.
Establish application-specific evals and define the quality, trust, latency, and cost thresholds required for production use.
Define the boundary and operating rhythm between the application team and the shared infrastructure and model teams.
Close the first critical hires for the applications team.
In the first 3 months
Launch a real product to a focused set of users, not just a demo or technology showcase.
Show evidence of repeated use and measurable customer value in the chosen workflow.
Establish instrumentation for activation, engagement, retention, quality, and cost to serve.
Demonstrate measurable user ROI, such as time saved, content found, decisions accelerated, or workflows completed.
Turn early customer-specific work into reusable workflows, templates, integrations, or application capabilities.
Translate application needs into a prioritized set of improvements for the core models, agent harness, and shared infrastructure.
Build a small team that can discover, build, launch, and iterate without heavy coordination overhead.
In the first 6 months
Establish at least one application with a credible path to product-market fit and durable distribution.
Turn early adoption into repeatable growth and monetization.
Create a clear portfolio thesis for what TwelveLabs should build next, based on what we have learned rather than top-down speculation.
Make the application layer a compounding advantage for the entire company by sharpening our platform, models, customer understanding, and go-to-market motion.
BACKGROUNDS THAT COULD FIT
There is no single required path. You may be:
A product leader who has repeatedly shipped zero-to-one products and stayed accountable for growth after launch.
A product growth leader with strong technical depth and a record of changing the product, not only optimizing funnels
A growth or product engineering leader who became the owner of a product and business outcome.
A founder or early startup leader who found a wedge, shipped the product, won the first users, and built the first team.
What matters is the combination: product taste, technical fluency, growth instinct, and the ability to build a team and business from a blank page.
Benefits and Perks
🤝 An open and inclusive culture and work environment
🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology
🏥 Full health, dental, and vision benefits
🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years
💪 Monthly wellness stipend
📚 Annual Learning & Development stipend to invest in your growth
💼 Global offices in San Francisco, New York, and Seoul, and coworking office memberships for remote team members
🛂 VISA support where applicable
🚆 Transportation stipend (SF & NY onsite only)
🍲 Daily meals provided (SF & NY onsite only)
","descriptionPlain":"WHO WE ARE\n\n\n\nVideo is 90% of the world's data. Most of it is invisible to machines.\n\nTwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.\n\nWe have raised over $200M from Silicon Valley's best VCs, including Index Ventures and NEA, and from leading enterprises like Nvidia and Amazon that are shaping the future of AI.\n\nWe are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!\n\n\n\nABOUT THE ROLE\n\n\n\nTwelveLabs spent years building the intelligence layer for video: shared infrastructure, multimodal video models, and an agent harness — the orchestration layer — that can understand and reason over video. Our vision for video superintelligence and pioneering work for large-scale video understanding allowed us to raise over $200M from Silicon Valley's best VCs like Index Ventures and NEA and leading enterprises like Nvidia and Amazon that are shaping the future of AI. The next chapter for TwelveLabs is to turn that foundation into products that people can use without needing to understand the technology underneath.\n\nWe are hiring a Head of Applications to build that layer.\n\nYou will decide where TwelveLabs should play, find the first use cases where our technology creates a step change in the user experience, and turn those insights into products customers use and pay for. You will own the application layer from customer discovery through product, engineering, launch, adoption, and growth.\n\nThis is a builder-GM role. It is not a conventional product management job, a growth marketing job, or a large-team executive role. You should be comfortable talking with customers in the morning, working through a product or technical tradeoff in the afternoon, and assembling an exceptional team for the 0-to-1 journey that evening. Ideal candidates will come with 12+ years of experience across product and product leadership roles, preferably at early to mid stage software companies.\n\n\n\nIN THIS ROLE, YOU WILL\n\n - Set the strategy for TwelveLabs' application layer: which users and workflows to serve, what to build, and what not to build.\n\n - Find a focused wedge through direct customer discovery, rapid prototyping, and disciplined market testing.\n\n - Ship opinionated, AI-native applications for non-technical users on top of TwelveLabs' core models, agent harness, and shared infrastructure. Define application-specific evals, task-success criteria, reliability thresholds, human review, and launch guardrails.\n\n - Own the full outcome of the applications you ship, including activation, engagement, retention, monetization, and customer value. Shipping is the start, not the finish.\n\n - Build a fast learning loop between users and the product. Turn usage data and customer behavior into better workflows, interfaces, and model experiences.\n\n - Act as the demanding internal customer for our platform and model teams. Define the capabilities, APIs, latency, quality, and reliability the application layer needs without rebuilding the layers below it.\n\n - Partner closely with Design, Research and Engineering, Product, Commercial, Revenue, and Partnerships to take products from prototype to repeatable adoption. Work deeply with lighthouse customers, while converting what you learn into reusable products rather than one-off implementations.\n\n - Recruit and lead a lean and extremely high-performing team with the range to move from idea to production.\n\n - Establish a culture of speed, craft, direct user contact, and clear accountability. Keep the team lean as the product and business grow.\n \n \n\nYou May be a good fit if you have\n\n - You have built and grown a product from zero to meaningful adoption. You can explain what users did, what changed because of your decisions, and where you were wrong.\n\n - You have exceptional product judgment and know what great looks like. You can find the narrow workflow that matters, reduce a complex technology to a simple experience, and say no to attractive distractions.\n\n - You are technically fluent enough to work directly with engineers and researchers on architecture, model behavior, latency, evaluation, reliability, and cost. You do not need to be the strongest engineer in the room, but you cannot treat the technology as a black box.\n\n - You understand the realities of AI-native products: probabilistic behavior, evaluation design, trust, human review, privacy and rights constraints, and the tradeoffs between quality, latency, and cost.\n\n - You understand growth as part of the product. You have designed or led loops that improved activation, retention, distribution, or monetization, not just acquisition campaigns.\n\n - You recruit well and have built an early team of unusually strong, low-ego people. You know how to operate before every function has its own leader.\n\n - You are close to the work. You use the product, talk with users, inspect the data, and raise the quality bar through direct involvement.\n\n - You move quickly without confusing activity with progress. You create evidence, make decisions, and change course when the evidence says you should.\n\n - You communicate clearly across research, engineering, design, and go-to-market teams and can resolve cross-layer tradeoffs without creating organizational drag.\n \n \n\nWHAT SUCCESS LOOKS LIKE IN THE FIRST 90 DAYS\n\n - Build a clear view of the most promising users, workflows, and distribution paths for the applications built.\n\n - Narrow the field to one or two application wedges using an explicit market-attractiveness and right-to-win filter, with clear reasons to reject the rest.\n\n - Put working prototypes in users' hands and establish a weekly build-measure-learn cadence.\n\n - Establish application-specific evals and define the quality, trust, latency, and cost thresholds required for production use.\n\n - Define the boundary and operating rhythm between the application team and the shared infrastructure and model teams.\n\n - Close the first critical hires for the applications team.\n\nIn the first 3 months\n\n - Launch a real product to a focused set of users, not just a demo or technology showcase.\n\n - Show evidence of repeated use and measurable customer value in the chosen workflow.\n\n - Establish instrumentation for activation, engagement, retention, quality, and cost to serve.\n\n - Demonstrate measurable user ROI, such as time saved, content found, decisions accelerated, or workflows completed.\n\n - Turn early customer-specific work into reusable workflows, templates, integrations, or application capabilities.\n\n - Translate application needs into a prioritized set of improvements for the core models, agent harness, and shared infrastructure.\n\n - Build a small team that can discover, build, launch, and iterate without heavy coordination overhead.\n\nIn the first 6 months\n\n - Establish at least one application with a credible path to product-market fit and durable distribution.\n\n - Turn early adoption into repeatable growth and monetization.\n\n - Create a clear portfolio thesis for what TwelveLabs should build next, based on what we have learned rather than top-down speculation.\n\n - Make the application layer a compounding advantage for the entire company by sharpening our platform, models, customer understanding, and go-to-market motion.\n\nBACKGROUNDS THAT COULD FIT\n\nThere is no single required path. You may be:\n\n - A product leader who has repeatedly shipped zero-to-one products and stayed accountable for growth after launch.\n\n - A product growth leader with strong technical depth and a record of changing the product, not only optimizing funnels\n\n - A growth or product engineering leader who became the owner of a product and business outcome.\n\n - A founder or early startup leader who found a wedge, shipped the product, won the first users, and built the first team.\n\nWhat matters is the combination: product taste, technical fluency, growth instinct, and the ability to build a team and business from a blank page.\n\n\n\nBenefits and Perks\n\n🤝 An open and inclusive culture and work environment\n\n🚀 Work closely with a collaborative, mission-driven team on cutting-edge AI technology\n\n🏥 Full health, dental, and vision benefits\n\n🌴 Extremely flexible PTO and parental leave policy. Office closed the week of Christmas and New Years\n\n💪 Monthly wellness stipend\n\n📚 Annual Learning & Development stipend to invest in your growth\n\n💼 Global offices in San Francisco, New York, and Seoul, and coworking office memberships for remote team members\n\n🛂 VISA support where applicable\n\n🚆 Transportation stipend (SF & NY onsite only)\n\n🍲 Daily meals provided (SF & NY onsite only)"},{"id":"75b92173-c0ca-4a91-84e5-c56515dbb813","title":"Senior Platform Engineer","department":"Tech","team":"Engineering ","employmentType":"FullTime","location":"Seoul, South Korea","secondaryLocations":[],"publishedAt":"2026-08-31T04:02:02.507+00:00","isListed":true,"isRemote":true,"workplaceType":"Hybrid","address":{"postalAddress":{"addressCountry":"South Korea","addressLocality":"Seoul"}},"jobUrl":"https://jobs.ashbyhq.com/twelve-labs/75b92173-c0ca-4a91-84e5-c56515dbb813","applyUrl":"https://jobs.ashbyhq.com/twelve-labs/75b92173-c0ca-4a91-84e5-c56515dbb813/application","descriptionHtml":"영상은 전 세계 데이터의 90%를 차지하지만, 그 대부분은 아직 기계가 이해할 수 없는 영역에 머물러 있습니다.
트웰브랩스(TwelveLabs)는 이 문제를 해결하기 위해 기계가 영상을 이해할 수 있도록 하는 AI 인프라를 구축합니다. 사람보다 더 정교하고 깊이 있게 영상을 이해하는 AI 모델을 통해 영상 속 시·청각 정보를 종합 분석하여 미디어·엔터테인먼트, 스포츠, 보안, 공공 분야 전반의 워크플로우에 적용하고 있습니다.
NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, 네이버벤처스, 한국투자파트너스, Quadrille Capital, Red Bull Ventures 등 글로벌 투자자들로부터 누적 3천억원 이상의 투자를 유치했습니다. 또한 Fei-Fei Li, Silvio Savarese, Alexandr Wang 등 세계적인 AI 리더들이 자문진으로 함께하고 있습니다.
트웰브랩스는 샌프란시스코, 서울, 뉴욕, 런던 네 개의 오피스를 거점으로 운영되는 글로벌 기업입니다. 핵심 연구개발은 서울에서 이루어지며, 전 세계 오피스의 구성원들이 이를 기반으로 다양한 산업과 시장에 영상 이해 기술을 적용해가고 있습니다. 세상의 복잡함을 이해하는 기술을 만드는 만큼, 서로 다른 배경과 관점을 가진 사람들이 모일 때 더 나은 제품을 만들 수 있다고 믿습니다.
아직 세상에 없는 답을 정의해보고 싶으신 분, 내 손으로 직접 변화를 만들어내고 싶어하시는 분을 찾고 있습니다. 트웰브랩스에서 세상을 이해하는 기술을 함께 만들어가세요.
TwelveLabs의 AI 플랫폼을 뒷받침하는 시스템을 구축하고 운영할 Senior Platform Engineer 를 찾고 있습니다. 특히 Platform 및 SRE에 대한 강한 관점과 경험을 가진 분을 기대합니다.
이 포지션은 매우 높은 수준의 Hands-on Engineering을 요구합니다. 인프라, 백엔드 시스템, Reliability Engineering 전반을 다루며, 실제 Production 시스템에 대해 상당한 수준의 Ownership을 갖게 됩니다.
서로 다른 엔지니어링 영역의 경계가 명확하지 않은 상황에서도 문제를 해결할 수 있어야 합니다. Production 장애가 발생했을 때 단순히 \"담당 팀\"에 문제를 전달하는 것이 아니라, 직접 문제를 조사하고 기존 가정을 검증하며 시스템 전반을 추적해 근본 원인을 찾고, 해결이 Production에 반영될 때까지 주도적으로 문제를 해결합니다.
또한 엔지니어들이 빠르게 개발하고 배포할 수 있도록 인프라와 내부 플랫폼을 설계하고 구축하는 동시에, Production 시스템의 Reliability, Observability, Performance 및 Operational Maturity를 지속적으로 개선합니다.
기술적인 문제를 깊이 파고드는 것을 즐기고, 시스템의 여러 레이어를 자유롭게 넘나들며 문제를 해결하고, 특정 인프라 영역을 관리하는 것보다 문제 자체에 Ownership을 갖고 해결하는 것을 선호하는 분이라면 잘 맞을 것입니다.
TwelveLabs의 AI 제품을 지원하는 확장 가능한 인프라 및 내부 플랫폼 설계·구축·운영
Incident Response부터 Root Cause Analysis 및 장기적인 개선까지 Production 시스템의 Reliability와 운영 상태에 대한 Ownership
빠르고 안전한 배포를 위한 CI/CD 구축 및 원활한 운영을 위한 Observability 시스템 개선
Terraform 등의 도구를 활용한 확장 가능한 CI/CD Pipeline 및 Infrastructure Automation 구축
Production 시스템의 Reliability, Performance, Scalability 및 Cost Efficiency 개선
Product 및 Engineering 팀이 서비스를 보다 효율적으로 개발하고 배포할 수 있도록 Platform 및 Tooling 개발
Backend Engineer, ML Engineer, Researcher, Product 팀과 긴밀하게 협업하여 시스템 요구사항을 이해하고 적절한 Platform Solution 구축
명확한 답이나 Ownership이 정해지지 않은 기술적 문제를 직접 주도하여 조사부터 구현 및 Production Rollout까지 책임
AWS, GCP, Azure 등 Cloud 환경에서 Production 시스템을 구축하고 운영한 경험
Linux, Networking, Container, Kubernetes에 대한 높은 수준의 경험
Terraform, Ansible 또는 유사한 Infrastructure as Code 도구 활용 경험
CI/CD 및 Deployment 시스템 설계 및 운영 경험
Observability, Monitoring, Logging, Alerting 및 Production Incident Response 경험
Python, Go, TypeScript 또는 기타 Backend-oriented Language를 활용한 프로그래밍 또는 Scripting 경험
Production 시스템의 여러 레이어에 걸친 복잡한 문제를 Debugging한 경험
Distributed Systems 및 Cloud Architecture에 대한 탄탄한 이해
증상을 일시적으로 완화하는 데 그치지 않고 문제의 Root Cause를 끝까지 파고드는 문제 해결 방식
Ownership의 경계가 명확하지 않은 상황에서도 효과적으로 문제를 해결할 수 있는 능력
Infrastructure, Platform Engineering, SRE, DevOps 또는 유사한 Engineering Role에서 7년 이상의 경력
여러 Engineering 팀이 사용하는 Developer Platform 또는 Infrastructure Platform 구축 경험
High Availability가 요구되는 B2B SaaS 제품을 대규모로 운영한 경험
높은 수준의 Security 및 Compliance 요구사항을 고려한 시스템 설계 경험
Backend Service Development 및 Distributed Systems 경험
System Performance, Reliability 및 Infrastructure Cost 사이의 Trade-off를 고려하여 의사결정한 경험
빠르게 성장하는 Startup 또는 높은 수준의 Ownership을 요구하는 환경에서 근무한 경험
영어로 커뮤니케이션할 수 있는 분
저희는 하나의 역할이나 카테고리에 정확히 들어맞는 사람을 찾고 있지는 않습니다. 가장 뛰어난 후보자는 탄탄한 Infrastructure 기반을 갖추고 있으면서도 Software를 직접 작성하고, Backend 시스템을 Debugging하며, Production Incident에 대응하고, 자신이 운영하는 시스템의 Reliability를 지속적으로 개선할 수 있는 사람입니다.
스스로를 Infrastructure Engineer, Platform Engineer, SRE 또는 DevOps Engineer라고 부를 수도 있습니다. 하지만 저희가 가장 중요하게 보는 것은 어떤 타이틀을 가지고 있는지가 아니라 문제를 어떻게 접근하고 해결하는가입니다. Production에서 문제가 발생했을 때, \"왜 이런 문제가 발생했는지 이해하고 싶은 사람\" 이 바로 저희가 찾는 사람입니다. 시스템을 깊이 파고들고, 기존의 가정을 의심하고, 여러 레이어에 걸쳐 근거를 따라가며 문제의 Root Cause를 찾아냅니다. 그리고 단순히 문제를 해결하는 데서 끝나지 않고, 같은 문제가 다시 발생할 가능성을 낮출 수 있도록 시스템 자체를 개선합니다. 이런 방식으로 문제를 해결하는 것이 즐거우신 분, TwelveLabs에서 찾는 분입니다.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.
No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.
Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.
We build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.
Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.
The Cognition Models team owns the models that turn video into structured understanding and reasoning: Pegasus, our video-language model, and Jockey Core, the reasoning LLM behind Jockey. In the model stack we sit between Perception Models (embeddings and retrieval) and the agent system — taking what's retrieved and producing structured understanding and the reasoning to act on it.
We focus on multimodal systems with high instruction-following capability and complex, hierarchically structured outputs. Our work spans training infrastructure from pre-training to RL, temporal segmentation and structured metadata extraction, large-scale inference and serving systems, data-curation and evaluation pipelines, and building Jockey Core. We ship products with real-world value rather than doing research in isolation, working as a goal-oriented, cross-functional team of ML researchers and engineers — using the most advanced compute in the world, including NVIDIA B300s, to accelerate the research-to-production cycle.
Pegasus is TwelveLabs' video-language model — it turns video into useful analysis by reasoning over visuals, speech, audio, and on-screen text. A key capability is Segment, our time-based metadata feature: instead of a broad question about a video, customers define the exact segment types they care about and the metadata fields they want back, and Pegasus finds the relevant start and end times and returns structured metadata for each segment — titles, summaries, topics, people, visual subjects, confidence, or domain-specific labels. This turns video into time-based, structured data that flows directly into search, archive, editing, compliance, or content-management workflows.
Make Pegasus training and research iteration as fast as possible by optimizing the training code end to end, from pre-training through fine-tuning and RL.
Profile and optimize the full training cycle: data loading, model execution, memory use, distributed communication, checkpointing, and evaluation.
Implement improvements in kernels, compilation, precision, and parallelism that measurably reduce experiment turnaround and time to target model quality.
Work with researchers to make new architectures and training methods efficient, using benchmarks to validate speed, correctness, convergence, and reproducibility.
Set the technical direction for training performance, lead critical implementation work, and mentor engineers through code and design reviews.
Deep hands-on experience developing large-scale model training code in Python and PyTorch, with strong ML fundamentals.
A record of measurable training speedups through profiling, GPU optimization, and multi-node distributed training.
Experience leading complex technical projects and mentoring engineers, with sound judgment across performance, model quality, and maintainability.
Experience with CUDA or Triton, compiler optimization, or low-precision training on modern accelerators.
Experience training and optimizing multimodal models at scale, including long-context or RL training and rollout workflows.
Master's or PhD in Machine Learning, Computer Science, or a related technical field.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
Video is 90% of the world's data. Most of it is invisible to machines.
TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government.
We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang.
We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us!
Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus.
No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product.
Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents.
We build on models we own. Marengo, our embedding model, resolves a query like \"the moment we almost missed the flight\" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end.
Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team.
This team develops two core capabilities for multimodal understanding: structure and semantics. Structure organizes video, audio, text, and documents into addressable units and models the relationships among them. Marengo, TwelveLabs' multimodal embedding model, represents what those units mean in a shared embedding space for understanding and retrieval.
End-to-end model development: We work across a broad range of research areas, including video segmentation and tracking, temporal and hierarchical modeling, contrastive learning, and multimodal representation learning. The team owns the entire model development lifecycle, from building large-scale training datasets and designing model architectures to optimizing distributed training and developing robust evaluation frameworks.
Research at scale: With access to world-class compute infrastructure, including NVIDIA B300 GPUs, we rapidly iterate on large-scale experiments, enabling fast progress on ambitious research problems.
Research with real-world impact: The path from research to production is exceptionally short. We work closely with the Agent, Search, Product, and Infrastructure teams to continuously improve the models that power multimodal search and understanding for thousands of customers worldwide.
As a Staff ML Research Engineer working across multimodal structure and embeddings, you will set the technical direction for TwelveLabs' next-generation models and own the end-to-end development process, from research strategy and data architecture to training systems, production model APIs, and evaluation frameworks.
This is a high-autonomy role at the intersection of video understanding, multimodal representation learning, large-scale systems design, and cross-team technical leadership. We're looking for someone who thrives in ambiguity: someone who can identify the highest-impact research problems, define the technical approach, and drive cross-team execution to deliver models that serve customers worldwide.
Set the technical direction for multimodal structure, including how assets are organized into reusable, addressable units and how those units relate.
Define the architecture, training, and data strategy for next-generation multimodal embedding models.
Own end-to-end model development from research planning through large-scale distributed training to production evaluation.
Architect and optimize large-scale training infrastructure, including distributed training pipelines, data processing systems, experiment workflows, and GPU utilization.
Own production model APIs end to end, from model packaging and API design to inference optimization and reliable operation at scale.
Drive data strategy by building large-scale curation, filtering, and quality systems for both structure and embeddings.
Define evaluation methods and quality standards for structure, embeddings, and end-to-end multimodal understanding and retrieval.
Define the interface between structure and semantics, ensuring that structured units remain reusable and addressable while improving end-to-end understanding and retrieval.
Drive cross-functional alignment with Agent, Search, Product, and Infrastructure teams on model integration and performance requirements.
Raise the research engineering bar through design review, experiment review, and technical mentorship.
Even if you don't check every box, we encourage you to apply.
If you're a zero-to-one achiever, a ferocious learner, and a kind team player who motivates others, you'll find a home at TwelveLabs.
7+ years of industry experience in computer vision, video understanding, or multimodal learning.
Demonstrated ability to take ambiguous, loosely-defined research problems and drive them to concrete, impactful solutions, from problem identification through delivery.
Strong judgment under changing constraints: you adapt as requirements evolve, make principled tradeoffs, and deliver the strongest result within the available time and resources.
Deep expertise in large-scale distributed model training (kernel optimization, FSDP, or similar).
Experience building and operating production model APIs for large-scale ML systems.
Deep expertise in video understanding, multimodal representation learning, or foundation model development.
Experience building end-to-end systems that connect multimodal structure with embeddings and retrieval.
Proven end-to-end ownership: not just running experiments, but defining what to build, building it, deploying it, and iterating on it in production.
Strong proficiency in Python and PyTorch.
Evidence of both research depth and engineering impact: publications paired with shipped products, not one or the other.
We evaluate based on relevant technical skills and sustained industry impact. This role is typically a strong fit for engineers with an MS and deep industry experience who have evolved from individual contributor to technical leader in production ML environments.
Experience training models at billion-parameter scale.
Experience with training operations: pipeline reliability, monitoring, fault tolerance, and cost optimization.
Experience with large-scale data curation and data quality systems.
Experience modeling temporal, spatial, or hierarchical structure across video and other multimodal content.
Experience with temporal video understanding or multimodal video modeling.
Deep experience optimizing training and inference systems for throughput, latency, GPU efficiency, and scale.
Track record of technical leadership: driving architectural decisions that shaped team or product direction.
The gap between research and production is remarkably short here. Models you build will be used by thousands of companies worldwide within months. In this role, you will shape how multimodal content is organized into reusable units and how those units are represented for understanding and retrieval. Rather than optimizing structure, embeddings, and retrieval in isolation, you will connect them into one end-to-end system. Our research philosophy balances rigorous experimentation with real-world application: we aim to build multimodal systems that are powerful, trustworthy, and genuinely useful.
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
영상은 전 세계 데이터의 90%를 차지하지만, 그 대부분은 아직 기계가 이해할 수 없는 영역에 머물러 있습니다.
트웰브랩스(TwelveLabs)는 이 문제를 해결하기 위해 기계가 영상을 이해할 수 있도록 하는 AI 인프라를 구축합니다. 사람보다 더 정교하고 깊이 있게 영상을 이해하는 AI 모델을 통해 영상 속 시·청각 정보를 종합 분석하여 미디어·엔터테인먼트, 스포츠, 보안, 공공 분야 전반의 워크플로우에 적용하고 있습니다.
NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, 네이버벤처스, 한국투자파트너스, Quadrille Capital, Red Bull Ventures 등 글로벌 투자자들로부터 누적 3천억원 이상의 투자를 유치했습니다. 또한 Fei-Fei Li, Silvio Savarese, Alexandr Wang 등 세계적인 AI 리더들이 자문진으로 함께하고 있습니다.
트웰브랩스는 샌프란시스코, 서울, 뉴욕, 런던 네 개의 오피스를 거점으로 운영되는 글로벌 기업입니다. 핵심 연구개발은 서울에서 이루어지며, 전 세계 오피스의 구성원들이 이를 기반으로 다양한 산업과 시장에 영상 이해 기술을 적용해가고 있습니다. 세상의 복잡함을 이해하는 기술을 만드는 만큼, 서로 다른 배경과 관점을 가진 사람들이 모일 때 더 나은 제품을 만들 수 있다고 믿습니다.
아직 세상에 없는 답을 정의해보고 싶으신 분, 내 손으로 직접 변화를 만들어내고 싶어하시는 분을 찾고 있습니다. 트웰브랩스에서 세상을 이해하는 기술을 함께 만들어가세요.
Jockey는 트웰브랩스의 통합 에이전틱 시스템으로, 영상과 이미지를 넘나들며 추론합니다. Reasoning Model에 메모리 레이어를 결합해, 사용자가 보유한 영상 데이터로부터 지식 저장소를 만들어냅니다.
아무리 큰 Context Window도 영상 아카이브 전체를 한 번에 담을 수는 없습니다. 저희가 다루는 영상은 수백만 시간 규모이기 때문입니다. 모델을 한 번 실행해서 얻는 답은 '영상 한 편'에 대한 것일 뿐, 수천~수만 개 영상으로 이루어진 영상 데이터 전체를 가로지르는 추론은 아닙니다 — Context Window를 아무리 늘려도 해결되지 않는 문제입니다. 그래서 Jockey는 사용자의 질문을 여러 단계로 쪼갠 뒤, 수천 개의 영상과 이미지에서 필요한 부분을 검색하고 구간을 나누어, 그 결과를 종합해 추론합니다. 예를 들어 아카이브를 지정하고 \"하이라이트 릴을 만들어줘\" 또는 \"가장 화제가 된 순간을 찾아줘\"라고 요청하면, 바로 사용할 수 있는 타임스탬프 구간을 결과로 돌려받습니다. 영상 데이터 전체를 이해하고, 그 이해를 실제 행동으로 옮길 수 있게 만드는 것 — 이것이 Jockey라는 제품의 핵심입니다.
사람뿐 아니라 에이전트를 위해 만듭니다. AI 에이전트가 영상의 주요 소비 주체로 점점 자리잡고 있습니다. 그래서 저희는 수백만 시간 규모로 확장하면서도, 사람과 자율 에이전트 모두에게 안정적이고 높은 품질의 결과를 제공하는 Production급 인프라를 구축하고 있습니다.
우리가 직접 보유한 모델 위에서 만듭니다. 임베딩 모델인 Marengo는 \"하마터면 비행기를 놓칠 뻔했던 순간\" 같은 쿼리를 실제 검색 결과로 풀어냅니다. 영상-언어 모델인 Pegasus는 사용자가 정의한 스키마에 따라 구조화된 타임스탬프 정보를 반환합니다. 두 모델을 지속적으로 출시하고 개선하기 때문에 고객은 별도의 재통합 없이도 릴리스를 거듭할수록 향상되는 Jockey의 품질을 누릴 수 있습니다. 엔드투엔드로 직접 통제하는 스택 위에서 에이전트를 만들 수 있는 팀은 많지 않습니다.
깊은 전문성, 하나의 시스템, 열린 문화. Foundation Model, Knowledge Construction, Search, Agent Harness가 모두 하나의 조직 안에 있습니다. 각 팀은 자신의 영역을 책임지고 그 안에서 깊은 전문성을 갖추되, 마치 F1 팀처럼 개별 부품이 아닌 전체 시스템 최적화를 지향합니다. 에이전트가 할 수 있는 일을 확장하지 못하는 모델 성능 향상은 성과로 보지 않습니다. 하나의 알고리즘 변경이 최종 시스템 동작에 미치는 영향까지 끝까지 추적하고, 완료된 결과물뿐 아니라 진행 중인 작업도 매주 공유합니다. 누구든 다른 팀으로부터 필요한 맥락을 자유롭게 가져올 수 있습니다.
트웰브랩스의 Construction (Video Ingestion & Serving Platform) 팀은 고객의 영상이 업로드된 순간부터 AI 모델이 검색, 분석, 생성에 활용할 수 있는 형태로 처리되고 저장되기까지의 End-to-End 플랫폼을 개발합니다. 영상 수집, 디코딩, 임베딩, 메타데이터 처리, 저장, 모델 서빙을 연결하며, 대규모 멀티테넌트 SaaS와 엔터프라이즈 배포 환경 모두에서 안정적으로 동작하는 시스템을 만듭니다.
Construction 팀은 Python과 Go 기반 서비스, Temporal, PostgreSQL, Kubernetes, KServe/vLLM, Terraform, ArgoCD, Grafana/OpenTelemetry 등을 활용합니다. Backend, Data, Infrastructure의 경계를 넘나들며 제품과 모델의 성장을 뒷받침합니다.
Video Ingestion & Serving Platform의 기술 방향과 프로덕션 운영을 주도할 Staff ML Engineer를 찾고 있습니다. 이 포지션은 장시간 실행되는 대규모 영상 처리 및 GPU 워크로드를 안정적으로 운영하고, 처리량, 지연 시간, 비용 효율성을 지속적으로 개선합니다. 또한 데이터 계층과 인프라를 포함한 주요 아키텍처와 마이그레이션을 안전하게 설계하고 실행합니다.
Staff Engineer로서 직접 코드를 작성하고 프로덕션 이슈를 해결하는 동시에, 여러 팀이 함께 활용할 수 있는 기술 기준과 플랫폼 역량을 만들어갑니다. 문제의 원인이 서비스, 데이터베이스, 워크플로우, Kubernetes 어느 곳에 있든 근거를 바탕으로 추적하고, 장기적으로 운영 가능한 해결책을 만드는 분을 기대합니다.
대규모 영상 처리, 임베딩 및 AI 모델 서빙을 위한 Backend Service와 Platform을 설계, 개발, 운영
Temporal 등 Durable Workflow 기반으로 장시간 실행되는 작업의 동시성, 재시도, Backpressure 및 장애 복구 체계 개선
Load Test, Profiling 및 운영 지표를 기반으로 처리량, 지연 시간, GPU 활용률과 인프라 비용 최적화
PostgreSQL 기반 데이터 모델, Sharding 및 Query Path를 설계하고 무중단 Production 운영 주도
Kubernetes, Terraform, ArgoCD 기반의 인프라와 CI/CD를 구축하고 Karpenter, KEDA 등을 활용한 확장성 개선
Metrics, Traces, Logs, Alerting을 포함한 Observability와 Incident Response 체계를 고도화하여 서비스 Reliability 향상
Backend, ML, Infrastructure, Product 팀과 협업하여 기술 방향을 정하고 주요 프로젝트의 Production Rollout 주도
소프트웨어 엔지니어링 경력 5년 이상 또는 이에 준하는 역량을 보유하신 분
Workflow, Queue, Database 등을 포함한 Distributed System 을 설계하고 Production 환경에서 운영한 경험이 있으신 분
Go 또는 Python 중 하나에 능숙하고 다른 언어로도 실무 개발이 가능하신 분
Kubernetes와 Cloud Infrastructure 위에서 서비스를 직접 운영하고 Terraform 등 IaC 도구로 자동화한 경험이 있으신 분
서비스 중단과 데이터 정합성을 고려하여 Database, Storage 또는 Backend 시스템의 Live Migration을 수행한 경험이 있으신 분
Load Test, Profiling, Monitoring을 통해 성능 병목을 찾고 처리량 또는 비용을 개선한 경험이 있으신 분
여러 팀에 걸친 기술적 의사결정을 주도하고, 설계 리뷰와 멘토링을 통해 팀의 실행력을 높인 경험이 있으신 분
빠르게 변화하는 환경에서 높은 수준의 Ownership을 바탕으로 문제 정의부터 Production 운영까지 주도할 수 있는 분
KServe, vLLM, Triton 등 GPU 기반 Inference Serving 시스템을 구축하거나 운영한 경험
Temporal, Cadence, Step Functions 등 Durable Workflow Engine을 활용한 경험
Aurora Limitless, Citus, Vitess 등 Sharded 또는 Distributed Database를 다룬 경험
Grafana, Mimir, Loki, Alloy, OpenTelemetry 기반 Observability Stack을 구축하고 운영한 경험
FFmpeg, 영상 디코딩, 트랜스코딩 또는 대규모 미디어 처리 파이프라인 경험
글로벌 팀과 영어로 원활하게 협업할 수 있는 커뮤니케이션 역량
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
영상은 전 세계 데이터의 90%를 차지하지만, 그 대부분은 아직 기계가 이해할 수 없는 영역에 머물러 있습니다.
트웰브랩스(TwelveLabs)는 이 문제를 해결하기 위해 기계가 영상을 이해할 수 있도록 하는 AI 인프라를 구축합니다. 사람보다 더 정교하고 깊이 있게 영상을 이해하는 AI 모델을 통해 영상 속 시·청각 정보를 종합 분석하여 미디어·엔터테인먼트, 스포츠, 보안, 공공 분야 전반의 워크플로우에 적용하고 있습니다.
NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, 네이버벤처스, 한국투자파트너스, Quadrille Capital, Red Bull Ventures 등 글로벌 투자자들로부터 누적 3천억원 이상의 투자를 유치했습니다. 또한 Fei-Fei Li, Silvio Savarese, Alexandr Wang 등 세계적인 AI 리더들이 자문진으로 함께하고 있습니다.
트웰브랩스는 샌프란시스코, 서울, 뉴욕, 런던 네 개의 오피스를 거점으로 운영되는 글로벌 기업입니다. 핵심 연구개발은 서울에서 이루어지며, 전 세계 오피스의 구성원들이 이를 기반으로 다양한 산업과 시장에 영상 이해 기술을 적용해가고 있습니다. 세상의 복잡함을 이해하는 기술을 만드는 만큼, 서로 다른 배경과 관점을 가진 사람들이 모일 때 더 나은 제품을 만들 수 있다고 믿습니다.
아직 세상에 없는 답을 정의해보고 싶으신 분, 내 손으로 직접 변화를 만들어내고 싶어하시는 분을 찾고 있습니다. 트웰브랩스에서 세상을 이해하는 기술을 함께 만들어가세요.
Jockey는 트웰브랩스의 통합 에이전틱 시스템으로, 영상과 이미지를 넘나들며 추론합니다. Reasoning Model에 메모리 레이어를 결합해, 사용자가 보유한 영상 데이터로부터 지식 저장소를 만들어냅니다.
아무리 큰 Context Window도 영상 아카이브 전체를 한 번에 담을 수는 없습니다. 저희가 다루는 영상은 수백만 시간 규모이기 때문입니다. 모델을 한 번 실행해서 얻는 답은 '영상 한 편'에 대한 것일 뿐, 수천~수만 개 영상으로 이루어진 영상 데이터 전체를 가로지르는 추론은 아닙니다 — Context Window를 아무리 늘려도 해결되지 않는 문제입니다. 그래서 Jockey는 사용자의 질문을 여러 단계로 쪼갠 뒤, 수천 개의 영상과 이미지에서 필요한 부분을 검색하고 구간을 나누어, 그 결과를 종합해 추론합니다. 예를 들어 아카이브를 지정하고 \"하이라이트 릴을 만들어줘\" 또는 \"가장 화제가 된 순간을 찾아줘\"라고 요청하면, 바로 사용할 수 있는 타임스탬프 구간을 결과로 돌려받습니다. 영상 데이터 전체를 이해하고, 그 이해를 실제 행동으로 옮길 수 있게 만드는 것 — 이것이 Jockey라는 제품의 핵심입니다.
사람뿐 아니라 에이전트를 위해 만듭니다. AI 에이전트가 영상의 주요 소비 주체로 점점 자리잡고 있습니다. 그래서 저희는 수백만 시간 규모로 확장하면서도, 사람과 자율 에이전트 모두에게 안정적이고 높은 품질의 결과를 제공하는 Production급 인프라를 구축하고 있습니다.
우리가 직접 보유한 모델 위에서 만듭니다. 임베딩 모델인 Marengo는 \"하마터면 비행기를 놓칠 뻔했던 순간\" 같은 쿼리를 실제 검색 결과로 풀어냅니다. 영상-언어 모델인 Pegasus는 사용자가 정의한 스키마에 따라 구조화된 타임스탬프 정보를 반환합니다. 두 모델을 지속적으로 출시하고 개선하기 때문에 고객은 별도의 재통합 없이도 릴리스를 거듭할수록 향상되는 Jockey의 품질을 누릴 수 있습니다. 엔드투엔드로 직접 통제하는 스택 위에서 에이전트를 만들 수 있는 팀은 많지 않습니다.
깊은 전문성, 하나의 시스템, 열린 문화. Foundation Model, Knowledge Construction, Search, Agent Harness가 모두 하나의 조직 안에 있습니다. 각 팀은 자신의 영역을 책임지고 그 안에서 깊은 전문성을 갖추되, 마치 F1 팀처럼 개별 부품이 아닌 전체 시스템 최적화를 지향합니다. 에이전트가 할 수 있는 일을 확장하지 못하는 모델 성능 향상은 성과로 보지 않습니다. 하나의 알고리즘 변경이 최종 시스템 동작에 미치는 영향까지 끝까지 추적하고, 완료된 결과물뿐 아니라 진행 중인 작업도 매주 공유합니다. 누구든 다른 팀으로부터 필요한 맥락을 자유롭게 가져올 수 있습니다.
트웰브랩스의 Construction (Video Ingestion & Serving Platform) 팀은 고객의 영상이 업로드된 순간부터 AI 모델이 검색, 분석, 생성에 활용할 수 있는 형태로 처리되고 저장되기까지의 End-to-End 플랫폼을 개발합니다. 영상 수집, 디코딩, 임베딩, 메타데이터 처리, 저장, 모델 서빙을 연결하며, 대규모 멀티테넌트 SaaS와 엔터프라이즈 배포 환경 모두에서 안정적으로 동작하는 시스템을 만듭니다.
Construction 팀은 Python과 Go 기반 서비스, Temporal, PostgreSQL, Kubernetes, KServe/vLLM, Terraform, ArgoCD, Grafana/OpenTelemetry 등을 활용합니다. Backend, Data, Infrastructure의 경계를 넘나들며 제품과 모델의 성장을 뒷받침합니다.
Construction 팀을 이끌고 Jockey Engineering 및 Research 조직 전반의 엔지니어링 실행 수준을 높일 Senior Engineering Manager를 찾고 있습니다.
이 포지션은 Construction 팀의 구성원, 실행, 기술 방향, 프로덕션 성과에 직접적인 책임을 집니다. 높은 성과를 내는 팀을 만들고, 우선순위를 명확히 하며, 엔지니어의 성장을 지원하고, 복잡한 플랫폼 프로젝트가 문제 정의부터 프로덕션 운영까지 안전하게 이어지도록 이끕니다.
기술적 깊이가 중요한 Engineering Manager 역할입니다. 주요 아키텍처와 운영 의사결정을 주도하고, 설계 및 코드 리뷰에 참여하며, 필요한 상황에서는 직접 문제 해결과 구현에 기여합니다. 문제의 원인이 서비스, 데이터베이스, 워크플로우, Kubernetes 또는 모델 서빙 인프라 어디에 있든 근거를 바탕으로 추적하고 팀이 어려운 프로덕션 문제를 해결하도록 이끌 수 있어야 합니다.
Construction 팀을 넘어 다른 Engineering Manager 및 Infrastructure, Backend, ML, Research 리더들과 협력하여 Jockey Engineering 및 Research 전반의 CI/CD, 릴리스 안정성, Observability, Incident Response, 개발 생산성, 테스트 및 AI 기반 개발 방식을 개선합니다. 초기의 전사적 영향력은 직접적인 조직 권한보다 기술적 신뢰, 협업, 그리고 다른 팀의 자발적인 도입을 통해 만들어갑니다.
명확한 목표, 피드백, 성과 관리, 채용, 커리어 개발 및 예측 가능한 실행 체계를 통해 Construction 팀을 리드하고 성장
Backend, ML, Infrastructure, Research 및 Product와 협력하여 대규모 영상 처리, 임베딩, 저장 및 AI 모델 서빙 플랫폼의 기술 방향 수립
동시성, 재시도, 멱등성, Backpressure, 장애 복구, PostgreSQL 확장 및 안전한 Live Migration을 포함한 Durable Workflow와 데이터 시스템 개선
Infrastructure 팀과 Kubernetes 및 IaC 기반 확장성을 발전시키고, Load Test, Profiling 및 운영 지표를 통해 처리량, 지연 시간, GPU 활용률, Reliability 및 비용 최적화
SLO, Observability, Alerting, On-call, Incident Response 및 Blameless Follow-up을 통해 운영 준비 수준 고도화
CI/CD, 자동화 테스트, 릴리스 및 Rollback 안정성, 개발 생산성, 서비스 Ownership, AI 기반 개발 표준을 개선하여 Jockey Engineering 및 Research 전반의 엔지니어링 효과성 향상
충분한 소프트웨어 엔지니어링 경력과 함께 프로덕션 시스템을 담당하는 팀을 구성하고 채용, 코칭, 성과 관리 및 커리어 개발을 수행한 경험
Workflow, Queue, Database, Storage 또는 대규모 데이터 처리를 포함한 Distributed System을 설계하고 운영하며 안전한 Live Migration을 이끈 경험
주요 코드 리뷰와 구현 의사결정을 주도하고 필요할 때 직접 기여할 수 있는 수준의 Go 또는 Python 역량
Kubernetes와 Cloud Infrastructure를 IaC로 운영하고 Load Test, Profiling, Monitoring 및 운영 데이터를 통해 성능, Reliability 또는 비용을 개선한 경험
보고 체계상의 권한에만 의존하지 않고 기술적 신뢰와 영향력을 바탕으로 여러 팀의 의사결정과 Production Rollout을 이끈 경험
모호한 문제를 정의하고 명확하게 소통하며 실행 속도와 Reliability, Maintainability 및 장기적인 Platform Health 사이에서 균형을 잡는 판단력
KServe, vLLM, Triton 등 GPU 기반 Inference Serving 시스템을 구축하거나 운영한 경험
Temporal, Cadence, Step Functions 등 Durable Workflow Engine을 활용한 경험
Aurora Limitless, Citus, Vitess, Cassandra, FoundationDB 등 Sharded 또는 Distributed Database를 다룬 경험
Grafana, Mimir, Loki, Alloy, OpenTelemetry 기반 Observability Stack을 구축하고 운영한 경험
FFmpeg, 영상 디코딩, 트랜스코딩 또는 대규모 미디어 처리 파이프라인 경험
여러 팀에 걸쳐 CI/CD, Developer Platform, 개발 생산성 또는 운영 방식을 개선한 경험
품질, 보안 및 유지보수 기준과 함께 AI 기반 개발 방식을 도입하고 운영한 경험
글로벌 팀과 영어로 원활하게 협업할 수 있는 커뮤니케이션 역량
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.
영상은 전 세계 데이터의 90%를 차지하지만, 그 대부분은 아직 기계가 이해할 수 없는 영역에 머물러 있습니다.
트웰브랩스(TwelveLabs)는 이 문제를 해결하기 위해 기계가 영상을 이해할 수 있도록 하는 AI 인프라를 구축합니다. 사람보다 더 정교하고 깊이 있게 영상을 이해하는 AI 모델을 통해 영상 속 시·청각 정보를 종합 분석하여 미디어·엔터테인먼트, 스포츠, 보안, 공공 분야 전반의 워크플로우에 적용하고 있습니다.
NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, 네이버벤처스, 한국투자파트너스, Quadrille Capital, Red Bull Ventures 등 글로벌 투자자들로부터 누적 3천억원 이상의 투자를 유치했습니다. 또한 Fei-Fei Li, Silvio Savarese, Alexandr Wang 등 세계적인 AI 리더들이 자문진으로 함께하고 있습니다.
트웰브랩스는 샌프란시스코, 서울, 뉴욕, 런던 네 개의 오피스를 거점으로 운영되는 글로벌 기업입니다. 핵심 연구개발은 서울에서 이루어지며, 전 세계 오피스의 구성원들이 이를 기반으로 다양한 산업과 시장에 영상 이해 기술을 적용해가고 있습니다. 세상의 복잡함을 이해하는 기술을 만드는 만큼, 서로 다른 배경과 관점을 가진 사람들이 모일 때 더 나은 제품을 만들 수 있다고 믿습니다.
아직 세상에 없는 답을 정의해보고 싶으신 분, 내 손으로 직접 변화를 만들어내고 싶어하시는 분을 찾고 있습니다. 트웰브랩스에서 세상을 이해하는 기술을 함께 만들어가세요.
트웰브랩스의 인프라 엔지니어는 AI SaaS 플랫폼을 안정적이고 확장 가능하게 운영할 수 있도록 핵심 인프라를 설계하고 구축합니다. 다양한 클라우드 환경과 온프레미스 환경에서의 시스템 아키텍처를 다루며, 영상 AI 파운데이션 모델을 뒷받침하는 견고한 인프라를 만들어 나갑니다. 빠르게 변화하는 스타트업 환경 속에서 성능, 보안, 유연성을 최적화하며, 사내 여러 팀과 긴밀히 협업하게 됩니다.
글로벌 기업 고객을 위한 멀티테넌트 아키텍처 설계 및 운영
Terraform을 통한 자동화와 확장성 높은 CI/CD 파이프라인 개발
AWS, GCP, Azure 등 다양한 클라우드 환경과 온프레미스를 아우르는 유연한 인프라 구축
고도화된 모니터링·보안 체계를 통해 안전하면서도 효율적인 클라우드 인프라 최적화
새로운 영상 AI 모델과 서비스를 가장 빠르게 지원하기 위한 확장 가능한 아키텍처 설계
PM, 엔지니어, 리서치와 협업하여 AI 제품을 현실로 구현
AWS, GCP, Azure 등 클라우드 환경에서 인프라 구축 및 운영 경험을 보유하신 분
Terraform, Ansible 등 IaC(Infrastructure as Code) 도구를 활용한 자동화 및 아키텍처 설계 경험이 있으신 분
Kubernetes, 컨테이너 기반 워크로드 운영 경험이 있으신 분
Python, Go, TypeScript 등 언어를 활용한 스크립팅 및 자동화 경험이 있으신 분
CI/CD 파이프라인 설계 및 운영 경험이 있으신 분
인프라 엔지니어로서 8년 이상의 경력을 보유하신 분
엔터프라이즈 SaaS 환경에서 멀티 테넌트 아키텍처 설계 및 운영 경험이 있으신 분
보안·컴플라이언스 표준에 맞춘 인프라 아키텍처 설계 및 감사 대응 경험이 있으신 분
고도화된 모니터링 및 로깅 시스템 구축 경험이 있으신 분
성능 최적화 및 비용 효율화를 동시에 고려한 인프라 운영 경험이 있으신 분
영어 커뮤니케이션이 가능하신 분
서류 검토 → 리크루터 인터뷰 → Hiring Manager 인터뷰 → 직무 인터뷰(1~2 Round) → 파이널 인터뷰 → 레퍼런스 체크 및 오퍼 → 최종합격
*전형절차는 직무별로 다르게 운영되며, 일정 및 상황에 따라 변동될 수 있습니다. 직무 인터뷰에는 대면 인터뷰가 포함될 수 있는 점 참고 바랍니다.
Growth & Tools
글로벌 B2B 고객과 함께 성장하는 Global Team
자율성과 협업을 모두 갖춘 하이브리드 근무
최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체
Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원
강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원
영어 교육 프로그램 및 글로벌 버디 프로그램 운영
야간 및 주말 출퇴근 택시비 지원
Meal & Snack
식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공
사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등)
사무실 근무 시, 오후 7시 이후 저녁 식대 제공
Wellness & Family
연 1회 본인 및 가족 1인의 건강검진 제공
단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1)
독감 예방접종비 지원
연말 2주간 유급 Holiday Break 운영
한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분
모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다.
채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다.