About the role
We are looking for a Research Scientist (Locomotion) with deep experience in reinforcement learning and real-world legged robot systems to join our team. You will be a foundational hire on a brand new initiative, building our locomotion stack from 0 to 1 and enabling dynamic, human-like behaviors on real robots. This is a high-ownership role at an early-stage, full-stack robotics company where you’ll work from simulation through to real-world execution.
What will you be doing?
- Train and iterate on RL-based locomotion policies for legged robots (bipedal, humanoid, quadruped)
- Drive sim-to-real transfer, bringing policies from simulators (Isaac Gym/Lab, MuJoCo, Genesis) onto real robot systems
- Design reward functions, curriculum learning strategies, and experiments to enable robust, agile behaviors (walking, recovery, dynamic motion, etc.)
- Work across the stack, from large-scale RL experimentation in simulation to testing and refining behaviors on physical robots
- Partner closely with a small, highly technical team to rapidly iterate between simulation and real-world systems and push forward a new robotics capability at the company
