Member of Technical Staff, Model Training (infra & platform)

What you'll do

You take technical ownership of the infrastructure that lets researchers train and deploy models. That means shared pipelines for supervised fine-tuning, reinforcement learning and on-policy distillation across language, multimodal and robotics workloads, with GPU workers, rollout generation, teacher-model inference, reward computation and simulation environments all coordinated.

Beyond the pipelines, the ad puts you in charge of the core architecture of micro1's agent ecosystem (execution runtimes, tool interfaces, model registries, serving APIs), plus experiment tracking, dataset and model versioning, checkpointing, fault recovery and observability. You also set technical standards across research, infrastructure and applied engineering, and mentor engineers while keeping hands-on ownership of the hard components. The role is full-time and remote.

What you need

• Strong computer science fundamentals, particularly distributed systems, concurrency, networking and performance engineering
• Extensive experience designing, building and operating scalable backend systems or distributed computing platforms
• Python plus at least one systems language such as C++, Rust or Go
• Hands-on experience with GPU-based ML infrastructure, distributed training, model inference or large-scale compute orchestration
• A solid grasp of compute utilization, memory efficiency, communication overhead, scheduling and fault tolerance
• Experience designing reliable APIs, services and infrastructure components
• Demonstrated technical ownership: architectural decisions, cross-functional leadership, systems other engineers depend on
• Comfort in fast-moving research settings where requirements keep changing

Nice to have

• PyTorch Distributed, DeepSpeed, Ray, Megatron-LM or similar
• GPU inference engines such as vLLM or TensorRT-LLM
• Reinforcement learning infrastructure, rollout orchestration or distributed agent execution
• CUDA, GPU kernel optimization or NCCL
• Kubernetes, cluster scheduling and cloud GPU infrastructure
• Simulation platforms, robotics infrastructure or large-scale agent evaluation systems
• Open-source contributions to distributed or high-performance ML systems

What it pays

The compensation notice inside the ad gives a national base salary range of $200,000–$320,000 for the full-time position, and that is the figure on this page, with equity for all employees and possible performance bonuses depending on role. micro1's card shows a wider $600,000–$1,800,000/year. The detail panel also shows an hourly figure of $7–8 that matches neither, so treat it as a display error and ask about the real package early. The notice lists up to 100% reimbursement of health-insurance premiums, paid time off and a 401(k) with company match, which makes this an employee-style package rather than the contractor engagement most micro1 listings carry.

Worth knowing

Good:

• A full-time role with a stated benefits package, unusual among micro1's listings
• Real architectural ownership over training and serving infrastructure, not ticket work
• Fully remote, with a scope that spans RL, distillation and agent runtimes
• The ad names concrete tools (vLLM, NCCL, Ray, Megatron-LM), so the bar is easy to gauge

Less good:

• The pay figures on the listing disagree with each other, so the real offer is unclear until you ask
• One opening, and the requirements read senior; this is a staff-level bar
• The ad does not say which time zones or hours the team keeps
• Hiring uses AI tools to assist screening, as the ad itself discloses

About this listing

Read on micro1's jobs board on 28 October 2026. The three Model Training listings (infra and platform, applied research, post-training research) share one pay line and one compensation notice but have different scopes; this page covers the infra and platform role. The compensation notice, benefits and every requirement above are the ad's published text. See micro1.