Senior Machine Learning Engineer, RL / Locomotion

What you'll do

• Design, train, and deploy reinforcement learning policies for legged robot locomotion using GPU-parallelized simulation.
• Develop terrain curriculum and domain randomization strategies to produce robust locomotion policies in real-world conditions.
• Own the sim-to-real transfer pipeline, identifying and closing gaps between simulated training and real-world deployment.
• Train policies for critical tasks such as stair climbing, rough terrain traversal, payload carry, push recovery, and fall recovery.
• Define and evaluate performance metrics to ensure locomotion robustness across various challenging terrains.

What we're looking for

• 3-8 years of experience with reinforcement learning for legged or mobile robots.
• Strong background in dynamics, control, and robot locomotion.
• Experience with RL algorithms (PPO, SAC) and training frameworks like rsl_rl, Stable Baselines.
• Hands-on experience with physics simulation tools such as Isaac Gym, MuJoCo, PyBullet.
• Proficiency in Python and PyTorch for developing machine learning models.