Senior Software Engineer, Agentic Robotics Infrastructure

• Master’s degree in Computer Science, Robotics, Engineering, or a related field (or equivalent experience).
• 5+ years of professional software engineering experience.
• Fluency across the kinds of software a robotics stack is made of - policy training, simulation, on-robot deployment - and the judgment to work effectively in whichever language and toolchain each of those calls for.
• Hands-on experience building and maintaining CI/CD pipelines (GitHub Actions, GitLab CI, or similar), with proven Docker and Linux fundamentals.
• Practical experience using coding agents or LLM-based developer tools in real work, and a point of view on where they succeed and where they fall over.
• A bias toward making other engineers faster, and the follow-through to get a practice embraced rather than just proposed.
• Clear written communication - a large part of this role is documentation, templates, and persuading teams to change how they work.

Ways to stand out from the crowd:

• Experience building agent skills, tool integrations, or MCP servers, and evaluating agent performance systematically.
• Familiarity with NVIDIA robotics products such as Jetson, Isaac Sim/Lab, or Isaac ROS.
• Background in robotics, embedded systems, or another domain where software has to run on real hardware.
• Experience managing GPU-backed CI infrastructure or self-hosted runner fleets.

Your base salary will be determined based on your location, experience, and the pay of employees in similar positions. The base salary range is 152,000 USD - 241,500 USD for Level 3, and 184,000 USD - 287,500 USD for Level 4.

Join NVIDIA’s Isaac Applications Engineering team and help build the platform for Physical AI robots - assembling the full robotics stack and proving it out on real hardware, from simulation through deployment. This role owns a question that runs through all of it: is our software ready to be used by agents? That means the coding agents we build with, the agents our customers will use to build on our stack, and the embedded agents running on the robot itself. You’ll design for it, build for it, test for it, and build the infrastructure that keeps it true as the stack evolves.

What you’ll be doing:

• Design for it. Shape APIs, error surfaces, tooling, and documentation so that agents can build with our robotics stack as effectively as human engineers - across multiple products and neighboring teams.
• Build for it. Create agent skills and automated workflows that teams adopt in their own repos, and support embedded agents on the robot (code-as-policy and related approaches) where latency, safety, and determinism constraints apply.
• Test for it. Run and integrate skill evaluations so agentic workflows are measured, not just deployed, and turn those evaluations into automated checks that catch regressions before they land.
• Build the infrastructure that keeps it true. Own CI/CD across the applications repos - spanning CMake, Bazel, and Python builds - including runners, Docker images, dataset provisioning, nightly builds, and release and docs pipelines.
• Spread what works. Package patterns into docs, templates, and starter kits, then drive adoption - running the conversations with teams and following through until a practice is live in more than one repo.