Run Robotics is building the debugging and evaluation layer for robot learning. We connect real world rollouts to the data, annotations, evals, and policy versions that produced them, so teams can understand why performance moved and what to fix next.
In this role, you will build the world model and simulation infrastructure that reproduces important failures, measures policy changes, and stays calibrated to physical robot outcomes. Your work will let researchers screen changes quickly while preserving the evidence behind every result.
What you’ll do
- Build evaluation environments that replay and perturb failures observed in real robot rollouts.
- Develop calibration methods that measure when a simulator or video world model tracks physical performance and when it does not.
- Create scalable evaluation harnesses for comparing policy checkpoints across task and failure distributions.
- Design metrics, datasets, and experiments that connect virtual scores to focused physical validation.
- Work directly with robotics teams to integrate their policies, rollouts, simulators, and world models.
What we’re looking for
- Strong software engineering ability in Python and modern ML tooling.
- Hands on experience with robot learning, simulation, video models, reinforcement learning, or embodied AI.
- Experience designing rigorous experiments and interpreting noisy evaluation results.
- Ability to own ambiguous research engineering problems from first prototype through reliable production infrastructure.
Nice to have
- Experience with MuJoCo, Isaac Sim, Genesis, ManiSkill, or custom simulation stacks.
- Research or production experience with world models, imitation learning, or vision language action policies.
- Experience calibrating offline or simulated metrics against real world behavior.
How we work
We are a small, technical team working closely with robotics researchers. We value direct evidence, clear writing, fast experiments, and the humility to update our view when the robot disagrees.
Run Robotics is an equal opportunity employer. We consider qualified applicants without regard to race, color, religion, sex, gender identity, sexual orientation, national origin, disability, veteran status, or any other protected characteristic.