Software Engineer, Simulation Platform

About the job

We're looking for a Software Engineer, Simulation Platform to build the software backbone that makes large-scale robot simulation fast, reproducible, and easy to use across the company.

This is not a 3D content or scene-authoring role. You will own the platform around simulation—the orchestration, infrastructure, APIs, and tooling that let autonomy, RL, and perception teams launch thousands of simulation runs, generate training data at scale, and trust the results. Your work turns simulation from a collection of one-off scripts into production infrastructure on the critical path to better models.

You'll work across distributed systems, cloud and GPU infrastructure, data pipelines, and developer tooling. You should be excited about building systems with tight feedback loops where throughput, reproducibility, and correctness directly shape how quickly our robots learn.

What you'll do

- Build the platform that runs simulation at scale—job orchestration, scheduling, and execution across cloud and GPU infrastructure

- Design APIs, SDKs, and tooling so autonomy, RL, and perception teams can define, launch, and reproduce simulation runs without fighting infrastructure

- Build the data engine that captures, versions, and serves synthetic datasets (RGB-D, segmentation, proprioception, contact, and related modalities) for training and evaluation

- Stand up benchmarking, regression, and CI harnesses so model and environment changes are evaluated automatically as the fleet and policies evolve

- Drive reproducibility: deterministic runs, asset and scene versioning, experiment tracking, and reliable replay

- Optimize throughput and cost across large simulation workloads (parallelization, batching, GPU utilization, caching)

- Integrate with simulation stacks (Isaac Sim / Isaac Lab, MuJoCo, Gazebo, or similar) behind clean platform abstractions

- Partner with simulation, RL, and deployment engineers to close the sim-to-real loop with better tooling, metrics, and observability

- Drive reliability, scalability, performance, and observability across a distributed, training-critical stack

What we're looking for

- Strong software engineering fundamentals and a track record of shipping production systems

- Experience in distributed systems, backend systems, cloud platforms, or large-scale batch/compute infrastructure

- Strong coding skills in Python, Go, C++, or similar

- Experience building data pipelines, job orchestration, or developer-facing tooling and APIs

- Ability to design systems that are robust, reproducible, and debuggable under real-world constraints

- Comfortable owning messy, cross-functional problems from architecture to execution

- High agency, strong judgment, and a bias toward building

Nice to have

- Experience with simulation stacks (NVIDIA Isaac Sim / Isaac Lab, MuJoCo, Gazebo, Unreal, Unity, or similar)

- Experience with GPU scheduling, ML training infrastructure, or experiment tracking platforms

- Experience building data pipelines or infrastructure for ML / AI training workflows at scale

- Familiarity with Docker, Kubernetes, cloud infrastructure, and event-driven or workflow systems

- Experience working with multimodal data such as telemetry, sensor streams, and video

- Exposure to robotics, RL, or sim-to-real workflows—enough to build the right abstractions for those teams

Who you are

- You want simulation to be a force multiplier for real robots, not a demo

- You care about throughput, reproducibility, and how good infrastructure compounds into faster learning

- You are ambitious, fast-moving, and highly technical

- You like hard problems, tight loops, and high ownership

- You want to help define how intelligent machines are trained at scale