Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for an Infrastructure Engineer to own the compute and cloud foundation that frontier AI research runs on. The systems you build determine how fast we can train models, how reliably experiments run, and how efficiently we scale. Some example areas you might work on:
Sandboxing and secure execution - design isolated environments where agents and untrusted code can run, use tools, and reach external services
Kubernetes and multi-cluster compute - operate CPU and GPU clusters as one platform with scheduling, autoscaling, and multi-tenant isolation
Training and inference infrastructure - understand the resource and scheduling demands of research, training, and inference workloads and build the platform capabilities those workloads need
Infrastructure for long-running agents - build the state management system to handle checkpointing, recovery, and resumption across failures
Networking - build the networking layer across clouds, clusters, and hosts: routing, peering, load balancing, and network isolation
If you're excited about building the infrastructure backbone of a frontier AI research lab - where your systems directly determine research velocity - we'd love to hear from you.
We offer a base salary of $300,000–$400,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for an Infrastructure Engineer to own the compute and cloud foundation that frontier AI research runs on. The systems you build determine how fast we can train models, how reliably experiments run, and how efficiently we scale. Some example areas you might work on:\n\n\n\n - Sandboxing and secure execution - design isolated environments where agents and untrusted code can run, use tools, and reach external services\n\n - Kubernetes and multi-cluster compute - operate CPU and GPU clusters as one platform with scheduling, autoscaling, and multi-tenant isolation\n\n - Training and inference infrastructure - understand the resource and scheduling demands of research, training, and inference workloads and build the platform capabilities those workloads need\n\n - Infrastructure for long-running agents - build the state management system to handle checkpointing, recovery, and resumption across failures\n\n - Networking - build the networking layer across clouds, clusters, and hosts: routing, peering, load balancing, and network isolation\n\n\n\n\nIf you're excited about building the infrastructure backbone of a frontier AI research lab - where your systems directly determine research velocity - we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"8cc02441-3d2e-451c-b39e-3b95ef8b20f3","title":"Member of Technical Staff, Pretraining","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-24T05:53:16.062+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/8cc02441-3d2e-451c-b39e-3b95ef8b20f3","applyUrl":"https://jobs.ashbyhq.com/mirendil/8cc02441-3d2e-451c-b39e-3b95ef8b20f3/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for an engineer to work at the intersection of research and systems on our pretraining stack. You'll contribute across the full pipeline, from data processing and model architecture to distributed training infrastructure and low-level optimization, and help determine how we scale our next generation of models. Some example areas you might work on (not limited to):
Implement and iterate on model architectures, training algorithms, and optimizer research in large-scale pretraining runs
Scale distributed training jobs across thousands of GPUs
Optimize training throughput for novel attention mechanisms, architecture variants, and compute efficiency improvements
Design and build large-scale data pipelines for efficient model consumption and dataset curation
Run and analyze scientific experiments to advance understanding of how architecture and data choices affect model capabilities
If you're excited about working across research and engineering to push the frontier of what large models can do, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for an engineer to work at the intersection of research and systems on our pretraining stack. You'll contribute across the full pipeline, from data processing and model architecture to distributed training infrastructure and low-level optimization, and help determine how we scale our next generation of models. Some example areas you might work on (not limited to):\n\n\n\n - Implement and iterate on model architectures, training algorithms, and optimizer research in large-scale pretraining runs\n\n - Scale distributed training jobs across thousands of GPUs\n\n - Optimize training throughput for novel attention mechanisms, architecture variants, and compute efficiency improvements\n\n - Design and build large-scale data pipelines for efficient model consumption and dataset curation\n\n - Run and analyze scientific experiments to advance understanding of how architecture and data choices affect model capabilities\n\n\n\nIf you're excited about working across research and engineering to push the frontier of what large models can do, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"7a0f4749-29a4-43c8-abf5-2f0269d53f7c","title":"Member of Technical Staff, Kernels","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-10-21T01:28:19.173+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/7a0f4749-29a4-43c8-abf5-2f0269d53f7c","applyUrl":"https://jobs.ashbyhq.com/mirendil/7a0f4749-29a4-43c8-abf5-2f0269d53f7c/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for an engineer to design, implement, and optimize custom ML kernels that bolster our model development stack. Your work will be deep in the system, combining hardware and software insights to optimize performance. Some example areas you might work on (not limited to):
Design and implement custom performant ML kernels that work at scale
Identify inefficiencies and optimize compute-intensive workloads to reduce memory bandwidth bottlenecks and improve hardware utilization
Enable and validate low-precision arithmetic formats and contribute to related compiler or runtime stacks
If you're excited about working at the intersection of hardware and frontier AI research, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for an engineer to design, implement, and optimize custom ML kernels that bolster our model development stack. Your work will be deep in the system, combining hardware and software insights to optimize performance. Some example areas you might work on (not limited to):\n\n - Design and implement custom performant ML kernels that work at scale\n\n - Identify inefficiencies and optimize compute-intensive workloads to reduce memory bandwidth bottlenecks and improve hardware utilization\n\n - Enable and validate low-precision arithmetic formats and contribute to related compiler or runtime stacks\n\n\n\nIf you're excited about working at the intersection of hardware and frontier AI research, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"235e6bac-aad6-4caa-9427-1aff7190c45b","title":"Member of Technical Staff, Post-Training, RL","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-24T05:55:09.424+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/235e6bac-aad6-4caa-9427-1aff7190c45b","applyUrl":"https://jobs.ashbyhq.com/mirendil/235e6bac-aad6-4caa-9427-1aff7190c45b/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for research engineers to help build the post-training stack for frontier reasoning models.
This role sits at the point where model capability, training dynamics, data, verification, and infrastructure all meet. You will design and run the experiments that turn a strong base model into a model that can solve difficult tasks reliably: choosing training objectives, shaping data mixtures, building verifiers, debugging reward signals, scaling runs, and understanding why a recipe works or fails.
Researchers are also expected to have strong engineering skills. The best work here will involve both: forming hypotheses about training behavior, implementing them in real systems, running large-scale experiments, reading the resulting traces carefully, and turning the lessons into the next training run.
Some areas you may work on include:
Post-training recipes: Develop and iterate on RL, SFT, and distillation recipes. Understand how choices in objectives, data mixtures, hyperparameters, rollout generation, and filtering affect efficiency, stability, capability, and final model behavior.
Scaling RL: Make post-training work at larger scales: more tokens, longer trajectories, larger models, more steps, and larger compute budgets. This includes identifying the bottlenecks that appear only when an approach leaves the small-run regime.
Long-horizon reasoning: Train models on tasks where success depends on many intermediate decisions. Develop methods for assigning useful feedback across long trajectories, where sparse rewards, credit assignment, exploration, and verification all become harder.
Off-policy and asynchronous training: Work on training regimes where data is generated by older policies, different policies, or partially filtered policies. Build intuition and tooling for when off-policy data helps, when it hurts, and how to control the resulting instabilities.
Verification and reward quality: Build robust verification pipelines for tasks where correctness can be checked automatically or semi-automatically. Detect and reduce reward hacking, false positives, brittle verifiers, and other failure modes that make RL look better than it really is.
Multi-task post-training: Scale recipes across different task families and domains. Study the tradeoffs between specialization and generality, and design training mixtures that improve all capabilities together.
Experiment analysis and debugging: Develop a deep empirical understanding of training runs. Diagnose regressions, separate real improvements from noise, design better ablations, and build the probes and analyses needed to make post-training less opaque.
End-to-end execution: Work closely with systems, infrastructure, and data teams to get experiments from idea to production-scale runs. This includes making training pipelines reliable, ensuring data and verifier quality, and turning successful experiments into repeatable and scalable recipes.
If you're excited about building the infrastructure that makes frontier RL research possible at scale, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for research engineers to help build the post-training stack for frontier reasoning models.\n\nThis role sits at the point where model capability, training dynamics, data, verification, and infrastructure all meet. You will design and run the experiments that turn a strong base model into a model that can solve difficult tasks reliably: choosing training objectives, shaping data mixtures, building verifiers, debugging reward signals, scaling runs, and understanding why a recipe works or fails.\n\nResearchers are also expected to have strong engineering skills. The best work here will involve both: forming hypotheses about training behavior, implementing them in real systems, running large-scale experiments, reading the resulting traces carefully, and turning the lessons into the next training run.\n\n\n\nSome areas you may work on include:\n\n - Post-training recipes: Develop and iterate on RL, SFT, and distillation recipes. Understand how choices in objectives, data mixtures, hyperparameters, rollout generation, and filtering affect efficiency, stability, capability, and final model behavior.\n\n - Scaling RL: Make post-training work at larger scales: more tokens, longer trajectories, larger models, more steps, and larger compute budgets. This includes identifying the bottlenecks that appear only when an approach leaves the small-run regime.\n\n - Long-horizon reasoning: Train models on tasks where success depends on many intermediate decisions. Develop methods for assigning useful feedback across long trajectories, where sparse rewards, credit assignment, exploration, and verification all become harder.\n\n - Off-policy and asynchronous training: Work on training regimes where data is generated by older policies, different policies, or partially filtered policies. Build intuition and tooling for when off-policy data helps, when it hurts, and how to control the resulting instabilities.\n\n - Verification and reward quality: Build robust verification pipelines for tasks where correctness can be checked automatically or semi-automatically. Detect and reduce reward hacking, false positives, brittle verifiers, and other failure modes that make RL look better than it really is.\n\n - Multi-task post-training: Scale recipes across different task families and domains. Study the tradeoffs between specialization and generality, and design training mixtures that improve all capabilities together.\n\n - Experiment analysis and debugging: Develop a deep empirical understanding of training runs. Diagnose regressions, separate real improvements from noise, design better ablations, and build the probes and analyses needed to make post-training less opaque.\n\n - End-to-end execution: Work closely with systems, infrastructure, and data teams to get experiments from idea to production-scale runs. This includes making training pipelines reliable, ensuring data and verifier quality, and turning successful experiments into repeatable and scalable recipes.\n\n\n\nIf you're excited about building the infrastructure that makes frontier RL research possible at scale, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"b56ba195-8baa-4f32-9891-20ef0cf6b18e","title":"Member of Technical Staff, Post-Training, RL Infra","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-24T05:56:33.468+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/b56ba195-8baa-4f32-9891-20ef0cf6b18e","applyUrl":"https://jobs.ashbyhq.com/mirendil/b56ba195-8baa-4f32-9891-20ef0cf6b18e/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for engineers to help build the post-training stack for frontier reasoning models. This role sits at the intersection of research and infrastructure. You will work to push the scale of our RL stack, whether it is novel recipe ideas, reliability, or performance. Some example areas you might work on (not limited to):
Design and build reliable infrastructure for large-scale RL training
Implement novel performance optimizations across the training stack
Develop evaluation and benchmarking infrastructure to measure model progress, throughput, and uptime
Build data collection and feedback pipelines that close the loop between human signal, reward modeling, and training
Collaborate with multiple teams to rapidly iterate on RL algorithms and get experiments into production training runs
If you're excited about building the infrastructure that makes frontier RL research possible at scale, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for engineers to help build the post-training stack for frontier reasoning models. This role sits at the intersection of research and infrastructure. You will work to push the scale of our RL stack, whether it is novel recipe ideas, reliability, or performance. Some example areas you might work on (not limited to):\n\n\n\n - Design and build reliable infrastructure for large-scale RL training\n\n - Implement novel performance optimizations across the training stack\n\n - Develop evaluation and benchmarking infrastructure to measure model progress, throughput, and uptime\n\n - Build data collection and feedback pipelines that close the loop between human signal, reward modeling, and training\n\n - Collaborate with multiple teams to rapidly iterate on RL algorithms and get experiments into production training runs\n\n\n\nIf you're excited about building the infrastructure that makes frontier RL research possible at scale, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"9fd66ea3-817c-481a-b5c5-357dc993e18c","title":"Member of Technical Staff, Post-Training, RL Environments","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-24T05:57:39.478+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/9fd66ea3-817c-481a-b5c5-357dc993e18c","applyUrl":"https://jobs.ashbyhq.com/mirendil/9fd66ea3-817c-481a-b5c5-357dc993e18c/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for a research engineer to build the data systems and execution environments that power reinforcement learning at Mirendil. The quality of our models depends directly on the quality of the data and environments we train on; you will own those systems end-to-end. Some example areas you might work on (not limited to):
Build and automate data collection pipelines for complex, long-horizon RL tasks.
Build robust systems to identify and prevent reward hacking.
Build scalable sandboxed execution environments for realistic tasks involving potentially multiple agents, nodes, and users.
Design systems to estimate the influence of training environments on production model behavior.
Collaborate with teams across the stack to identify potential axes of improvements in production model behavior, and develop training environments to push these axes.
If you're excited about building the data and environment infrastructure that determine what our models learn, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for a research engineer to build the data systems and execution environments that power reinforcement learning at Mirendil. The quality of our models depends directly on the quality of the data and environments we train on; you will own those systems end-to-end. Some example areas you might work on (not limited to):\n\n\n\n - Build and automate data collection pipelines for complex, long-horizon RL tasks.\n\n - Build robust systems to identify and prevent reward hacking.\n\n - Build scalable sandboxed execution environments for realistic tasks involving potentially multiple agents, nodes, and users.\n\n - Design systems to estimate the influence of training environments on production model behavior.\n\n - Collaborate with teams across the stack to identify potential axes of improvements in production model behavior, and develop training environments to push these axes.\n\n\n\nIf you're excited about building the data and environment infrastructure that determine what our models learn, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"2c7c136a-b24b-456f-afb3-d5530bad84b5","title":"Member of Technical Staff, Model Evaluation","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-24T05:58:56.441+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/2c7c136a-b24b-456f-afb3-d5530bad84b5","applyUrl":"https://jobs.ashbyhq.com/mirendil/2c7c136a-b24b-456f-afb3-d5530bad84b5/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for a research engineer to build the evaluation infrastructure that tells us whether our models are getting better in ways we care about. You'll own the frameworks, pipelines, and tooling that measure model behavior across capabilities. Some example areas you might work on (not limited to):
Design and build evaluation frameworks that measure model capabilities along realistic axes, beyond standard benchmarks.
Build automated eval pipelines and regression-detection systems that run continuously and surface signal quickly.
Develop agent-assisted workflows for humans to efficiently inspect model behavior.
Instrument training runs with observability tooling so researchers can understand what's changing in model behavior, and why.
Partner with post-training and RL teams to close the loop between eval signal and training decisions.
If you're excited about the hard problem of knowing whether a frontier AI system is actually improving, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for a research engineer to build the evaluation infrastructure that tells us whether our models are getting better in ways we care about. You'll own the frameworks, pipelines, and tooling that measure model behavior across capabilities. Some example areas you might work on (not limited to):\n\n\n\n - Design and build evaluation frameworks that measure model capabilities along realistic axes, beyond standard benchmarks.\n\n - Build automated eval pipelines and regression-detection systems that run continuously and surface signal quickly.\n\n - Develop agent-assisted workflows for humans to efficiently inspect model behavior.\n\n - Instrument training runs with observability tooling so researchers can understand what's changing in model behavior, and why.\n\n - Partner with post-training and RL teams to close the loop between eval signal and training decisions.\n\n\n\nIf you're excited about the hard problem of knowing whether a frontier AI system is actually improving, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"27f9eced-d320-4483-9d31-7e2fefb02c4c","title":"Member of Technical Staff, Inference","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-24T06:02:09.356+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/27f9eced-d320-4483-9d31-7e2fefb02c4c","applyUrl":"https://jobs.ashbyhq.com/mirendil/27f9eced-d320-4483-9d31-7e2fefb02c4c/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We are looking for an engineer to own the inference systems that power our models in production and research. You'll work across the full inference stack, from serving infrastructure down to hardware-level optimization. Some example areas you might work on (not limited to):
Design and build high-throughput, low-latency inference serving systems for frontier models, optimizing for both research iteration and production deployment
Optimize inference performance across GPU and accelerator hardware - maximizing FLOPs utilization, memory bandwidth, and compute efficiency for large-scale models
Enable and extend distributed inference frameworks (e.g. vLLM, SGLang, TensorRT-LLM) to support novel architectures, long-context workloads, and agentic inference patterns
Implement and validate inference-time optimizations: speculative decoding, quantization, KV cache management, and batching strategies
Build observability and reliability infrastructure so the team can measure latency, throughput, and cost across every serving configuration
Partner directly with teams to bring new model architectures and post-training techniques into production quickly
If you're excited about pushing the performance limits of frontier model inference, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe are looking for an engineer to own the inference systems that power our models in production and research. You'll work across the full inference stack, from serving infrastructure down to hardware-level optimization. Some example areas you might work on (not limited to):\n\n\n\n - Design and build high-throughput, low-latency inference serving systems for frontier models, optimizing for both research iteration and production deployment\n\n - Optimize inference performance across GPU and accelerator hardware - maximizing FLOPs utilization, memory bandwidth, and compute efficiency for large-scale models\n\n - Enable and extend distributed inference frameworks (e.g. vLLM, SGLang, TensorRT-LLM) to support novel architectures, long-context workloads, and agentic inference patterns\n\n - Implement and validate inference-time optimizations: speculative decoding, quantization, KV cache management, and batching strategies\n\n - Build observability and reliability infrastructure so the team can measure latency, throughput, and cost across every serving configuration\n\n - Partner directly with teams to bring new model architectures and post-training techniques into production quickly\n\n\n\nIf you're excited about pushing the performance limits of frontier model inference, we'd love to hear from you.\n\n\n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"24d91f89-d315-4c1b-bf84-dbe7314e30f4","title":"Member of Technical Staff, Recursive Self-Improvement (RSI)","department":"Tech","team":"Tech","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-06-12T04:03:55.888+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/24d91f89-d315-4c1b-bf84-dbe7314e30f4","applyUrl":"https://jobs.ashbyhq.com/mirendil/24d91f89-d315-4c1b-bf84-dbe7314e30f4/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
We are looking for an innovative, rigorous Research Engineer to join our team to accelerate AI self-improvement. This role requires a deep understanding of ML at both the application and system levels. You will ship AI-driven systems that recursively improve how AI systems are trained, evaluated, deployed, and operated at scale. If you are driven by the compounding potential of accelerating the AI loop, you will thrive in this role. Areas you might work on include:
Build autonomous AI that improves AI. Develop models, harnesses, and pipelines that automate parts of the ML lifecycle - data curation, training optimization, debugging, model selection, and experiment execution - and measure their impact on the speed and reliability of AI R&D.
Close loops across the stack. Identify the highest-leverage improvement opportunities anywhere in the stack, from distributed pre-training, post-training, and serving to agent harnesses, runtimes, and research environments, and jointly design and optimize across layers.
Run experiments end-to-end. Form hypotheses, design experiments, build the infrastructure to run them at scale, and turn results into shipped improvements.
Develop evaluation and observability. Build benchmarks, automated evals, and monitoring systems that surface regressions, failure modes, and emergent behaviors in AI systems.
If you're excited about closing the loop at scale, we'd love to hear from you.
We offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n \n\n\nTHE ROLE\n\nWe are looking for an innovative, rigorous Research Engineer to join our team to accelerate AI self-improvement. This role requires a deep understanding of ML at both the application and system levels. You will ship AI-driven systems that recursively improve how AI systems are trained, evaluated, deployed, and operated at scale. If you are driven by the compounding potential of accelerating the AI loop, you will thrive in this role. Areas you might work on include:\n\n - Build autonomous AI that improves AI. Develop models, harnesses, and pipelines that automate parts of the ML lifecycle - data curation, training optimization, debugging, model selection, and experiment execution - and measure their impact on the speed and reliability of AI R&D.\n\n - Close loops across the stack. Identify the highest-leverage improvement opportunities anywhere in the stack, from distributed pre-training, post-training, and serving to agent harnesses, runtimes, and research environments, and jointly design and optimize across layers.\n\n - Run experiments end-to-end. Form hypotheses, design experiments, build the infrastructure to run them at scale, and turn results into shipped improvements.\n\n - Develop evaluation and observability. Build benchmarks, automated evals, and monitoring systems that surface regressions, failure modes, and emergent behaviors in AI systems.\n\nIf you're excited about closing the loop at scale, we'd love to hear from you.\n\n \n\nWe offer a base salary of $300,000–$400,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"b15b1e15-aa73-4bed-8977-7e72e2443221","title":"Member of Technical Staff — Company Building","department":"Operations","team":"Operations","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-10-21T01:26:34.751+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/b15b1e15-aa73-4bed-8977-7e72e2443221","applyUrl":"https://jobs.ashbyhq.com/mirendil/b15b1e15-aa73-4bed-8977-7e72e2443221/application","descriptionHtml":"Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
Every member of technical staff at Mirendil pushes the frontier. Most do it through research and engineering. This role does it by building the company itself.
You'll operate as a force multiplier for the founders, sitting at the center of a frontier AI lab, owning founder priorities, ways of working, early customers, and special projects. The scope shifts with the company's priorities; the job is always the highest-leverage unowned problem.
Your technical background is not a nice-to-have here. Our customers are researchers and engineers. Our product is frontier AI R&D itself. You'll speak fluently with ML researchers in the morning, translate that work into a compelling narrative for a design partner at noon, and improve our research organization's ways of working in the afternoon. You'll work directly with the CEO and co-founders on the highest-leverage problems in the company: you'll see every part of how a frontier AI company is built, funded, positioned, and scaled, from the inside.
What You'll Own
Founder leverage. Drive the founders' highest-priority initiatives while providing the day-to-day operational support that keeps them focused on what matters most: preparing for critical partner and investor conversations, coordinating priorities, and turning decisions into execution. You'll keep the founders on track and make sure their commitments don't slip.
Zero to one with our first customers. Take our first customers from zero to one, then work with Business and Operations to scale the motions. Our buyers are highly technical; you will shape our positioning, land early design partners, and run pilots end-to-end.
Company operating rhythm. Own the company operating rhythm. Build and run the systems that keep a research-driven company aligned and fast: planning cycles, goal tracking and internal KPIs, internal comms, and cross-team execution. You're the person who makes sure decisions get made, tracked, and followed through on across the team.
Special projects. Jump on whatever is most important and least owned: a partnership negotiation, a launch, a market or technical landscape analysis, an operational fire.
Who You Are
Technically credentialed and technically credible. MS or PhD in CS, EE, physics, math, or another STEM field, plus a few years in industry as an engineer, scientist, or researcher. You can read a paper and credibly participate in an ML systems discussion, and hold your own with frontier researchers.
Pulled toward the business side. You've felt the itch beyond your IC scope: maybe you drove a cross-functional initiative, ran the customer-facing side of a technical project, started something of your own, or found yourself doing the strategy work nobody asked you to do.
A translator. Exceptional written and verbal communicator who can distill technical depth and ambiguity into crisp decisions and narratives for investors, partners, and customers.
Structured and fast. You bring the analytical rigor of your technical training to messy, open-ended business problems: decomposing them, prioritizing ruthlessly, and executing fast.
Extremely high energy and endurance. This is an intense, all-in role with long hours at a company moving at frontier-lab pace. You thrive on that, and you're at your best when the stakes and tempo are high.
Bias toward action and low ego. You default to doing, you're relentlessly resourceful, and nothing is beneath you or above you.
We offer a base salary of $200,000–$350,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"FOR ENGINEERS AND SCIENTISTS READY TO BUILD THE COMPANY, NOT JUST THE TECHNOLOGY\n\n\n\n\nMIRENDIL\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nEvery member of technical staff at Mirendil pushes the frontier. Most do it through research and engineering. This role does it by building the company itself.\n\nYou'll operate as a force multiplier for the founders, sitting at the center of a frontier AI lab, owning founder priorities, ways of working, early customers, and special projects. The scope shifts with the company's priorities; the job is always the highest-leverage unowned problem.\n\nYour technical background is not a nice-to-have here. Our customers are researchers and engineers. Our product is frontier AI R&D itself. You'll speak fluently with ML researchers in the morning, translate that work into a compelling narrative for a design partner at noon, and improve our research organization's ways of working in the afternoon. You'll work directly with the CEO and co-founders on the highest-leverage problems in the company: you'll see every part of how a frontier AI company is built, funded, positioned, and scaled, from the inside.\n\n\n\nWhat You'll Own\n\n - Founder leverage. Drive the founders' highest-priority initiatives while providing the day-to-day operational support that keeps them focused on what matters most: preparing for critical partner and investor conversations, coordinating priorities, and turning decisions into execution. You'll keep the founders on track and make sure their commitments don't slip.\n\n - Zero to one with our first customers. Take our first customers from zero to one, then work with Business and Operations to scale the motions. Our buyers are highly technical; you will shape our positioning, land early design partners, and run pilots end-to-end.\n\n - Company operating rhythm. Own the company operating rhythm. Build and run the systems that keep a research-driven company aligned and fast: planning cycles, goal tracking and internal KPIs, internal comms, and cross-team execution. You're the person who makes sure decisions get made, tracked, and followed through on across the team.\n\n - Special projects. Jump on whatever is most important and least owned: a partnership negotiation, a launch, a market or technical landscape analysis, an operational fire.\n\n\n\nWho You Are\n\n - Technically credentialed and technically credible. MS or PhD in CS, EE, physics, math, or another STEM field, plus a few years in industry as an engineer, scientist, or researcher. You can read a paper and credibly participate in an ML systems discussion, and hold your own with frontier researchers.\n\n - Pulled toward the business side. You've felt the itch beyond your IC scope: maybe you drove a cross-functional initiative, ran the customer-facing side of a technical project, started something of your own, or found yourself doing the strategy work nobody asked you to do.\n\n - A translator. Exceptional written and verbal communicator who can distill technical depth and ambiguity into crisp decisions and narratives for investors, partners, and customers.\n\n - Structured and fast. You bring the analytical rigor of your technical training to messy, open-ended business problems: decomposing them, prioritizing ruthlessly, and executing fast.\n\n - Extremely high energy and endurance. This is an intense, all-in role with long hours at a company moving at frontier-lab pace. You thrive on that, and you're at your best when the stakes and tempo are high.\n\n - Bias toward action and low ego. You default to doing, you're relentlessly resourceful, and nothing is beneath you or above you.\n \n \n\nWe offer a base salary of $200,000–$350,000 USD and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"9331140c-f137-4fdb-8524-f03ec60fea80","title":"Executive Operations ","department":"Operations","team":"Operations","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-10-21T01:27:38.546+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/9331140c-f137-4fdb-8524-f03ec60fea80","applyUrl":"https://jobs.ashbyhq.com/mirendil/9331140c-f137-4fdb-8524-f03ec60fea80/application","descriptionHtml":"Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
You are the operating system for the CEO. You'll keep the CEO's days tight and their commitments on track. Your scope will include facilities, vendors, purchasing and bill pay, team logistics, occasional events and travel, and running the CEO calendar, meeting logistics and travel support. You are the force that makes things feel seamless even when they aren’t.
What You'll Own
CEO leverage. Own the CEO’s calendars, travel, expenses, and inbox triage. Prep them for what's next, capture decisions, and turn \"we should\" into \"it's handled.\" You keep their days focused on the highest-value work and make sure nothing slips through.
The office runs like a product. Facilities, vendors, supplies, catering and snacks,seating, security, mail and shipping, the space itself. When something breaks, you fix it before anyone notices.
Fuel the team. Long, in-person days mean meals and daily rhythm are load-bearing. Own catering, snacks, coffee, and the small rituals that keep a high-intensity team fed and energized from morning to night.
Moments that matter. Team dinners, offsites, onboardings and the visitor experience for candidates, partners, and investors. Everyone who walks through the door should feel cared for.
Whatever's least owned. Jump on the most important unowned thing, whatever it is that week: an event thrown together on short notice, a move or buildout, an operational fire.
Ad-hoc support. It might be billing, it might be people ops, there’s plenty to do and space to grow.
Who You Are
You’ve been someone’s trusted right hand. An executive, founder, or small team has relied on you to handle important work independently. Exceptional EAs and candidates with broader operational backgrounds are welcome.
You exercise judgment, not just follow instructions. You can work with incomplete information, distinguish what matters from what doesn’t, and know when to act and when to escalate.
Your follow-through is exceptional. You keep multiple moving pieces organized, hold vendors accountable, and are simply on top of it all.
You combine high standards with low ego. You’re comfortable negotiating with a vendor and picking up a team lunch. You genuinely enjoy making other people more effective and comfortable.
Note, this role will be in person 5 days a week.
We offer a base salary of $150,000–$250,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"MIRENDIL\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nYou are the operating system for the CEO. You'll keep the CEO's days tight and their commitments on track. Your scope will include facilities, vendors, purchasing and bill pay, team logistics, occasional events and travel, and running the CEO calendar, meeting logistics and travel support. You are the force that makes things feel seamless even when they aren’t.\n\n\n\nWhat You'll Own\n\n - CEO leverage. Own the CEO’s calendars, travel, expenses, and inbox triage. Prep them for what's next, capture decisions, and turn \"we should\" into \"it's handled.\" You keep their days focused on the highest-value work and make sure nothing slips through.\n\n - The office runs like a product. Facilities, vendors, supplies, catering and snacks,seating, security, mail and shipping, the space itself. When something breaks, you fix it before anyone notices.\n\n - Fuel the team. Long, in-person days mean meals and daily rhythm are load-bearing. Own catering, snacks, coffee, and the small rituals that keep a high-intensity team fed and energized from morning to night.\n\n - Moments that matter. Team dinners, offsites, onboardings and the visitor experience for candidates, partners, and investors. Everyone who walks through the door should feel cared for.\n\n - Whatever's least owned. Jump on the most important unowned thing, whatever it is that week: an event thrown together on short notice, a move or buildout, an operational fire.\n\n - Ad-hoc support. It might be billing, it might be people ops, there’s plenty to do and space to grow.\n\n\n\nWho You Are\n\n - You’ve been someone’s trusted right hand. An executive, founder, or small team has relied on you to handle important work independently. Exceptional EAs and candidates with broader operational backgrounds are welcome.\n\n - You exercise judgment, not just follow instructions. You can work with incomplete information, distinguish what matters from what doesn’t, and know when to act and when to escalate.\n\n - Your follow-through is exceptional. You keep multiple moving pieces organized, hold vendors accountable, and are simply on top of it all.\n\n - You combine high standards with low ego. You’re comfortable negotiating with a vendor and picking up a team lunch. You genuinely enjoy making other people more effective and comfortable.\n\nNote, this role will be in person 5 days a week.\n\nWe offer a base salary of $150,000–$250,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits."},{"id":"d1ca71be-65e3-4400-977d-e486318b81b9","title":"Member of Technical Staff, Distributed Systems Engineer","department":"Tech","team":"ML Systems","employmentType":"FullTime","location":"San Francisco","secondaryLocations":[],"publishedAt":"2026-09-29T03:43:13.726+00:00","isListed":true,"isRemote":false,"workplaceType":"OnSite","address":{"postalAddress":{"addressCountry":"","addressLocality":"United States"}},"jobUrl":"https://jobs.ashbyhq.com/mirendil/d1ca71be-65e3-4400-977d-e486318b81b9","applyUrl":"https://jobs.ashbyhq.com/mirendil/d1ca71be-65e3-4400-977d-e486318b81b9/application","descriptionHtml":"Mirendil
Mirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.
The Role
We have a lot of compute, and a lot more work for it. We’re looking for a Systems Engineer who knows how to find bottlenecks and move them somewhere we can live with. Your work determines how much goodput we get out of our compute. Some example areas you might work on:
AI workload and system co-design - rethink the AI application, execution, and hardware choices together to improve performance across the stack
Performance engineering - profile across the stack, find where time and resources go, and fix what matters
Reliable distributed systems - design how work and data move across machines, balancing parallelism, communication, and correctness under load and failure
Heterogeneous compute and workloads - make large-scale mixes of sandboxes, agent harnesses, training, and inference run efficiently together across CPUs and accelerators
We offer a base salary of $300,000–$400,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.
","descriptionPlain":"Mirendil\n\nMirendil is a tech-first company focused on solving core bottlenecks that unlock step-change acceleration across science and technology. Our first goal is to democratize frontier AI R&D across scientific disciplines. We are building a frontier AI research company and training our own models end-to-end.\n\n\n\nThe Role\n\nWe have a lot of compute, and a lot more work for it. We’re looking for a Systems Engineer who knows how to find bottlenecks and move them somewhere we can live with. Your work determines how much goodput we get out of our compute. Some example areas you might work on:\n\n - AI workload and system co-design - rethink the AI application, execution, and hardware choices together to improve performance across the stack\n\n - Performance engineering - profile across the stack, find where time and resources go, and fix what matters\n\n - Reliable distributed systems - design how work and data move across machines, balancing parallelism, communication, and correctness under load and failure\n\n - Heterogeneous compute and workloads - make large-scale mixes of sandboxes, agent harnesses, training, and inference run efficiently together across CPUs and accelerators\n\n\n\nWe offer a base salary of $300,000–$400,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits."}],"apiVersion":"1"}