Senior Applied Scientist, Safe Locomotion, Compass
Core
Develop safe legged locomotion algorithms and learning-based controllers for quadrupeds and humanoids to walk, run, and recover from disturbances on physical hardware.
Role type
Senior Applied Scientist (Safe Locomotion)
Builds
Learning-based controllers and whole-body control laws for legged robots
Domain
Robotics, Machine Learning, Control Theory
Deliverable
production ML models
Required skills
Reinforcement Learning, sim-to-real transfer, legged robot dynamics, whole-body control, Python, deep learning frameworks, physics simulators
Preferred skills
Safety-constrained RL, model-based control, stability theory, hierarchical RL, real-time deployment constraints, large-scale training infrastructure
Technologies
PyTorch, JAX, Isaac Gym/Sim, MuJoCo, PyBullet
Responsibilities
Design and deploy RL policies for dynamic locomotion; integrate learned policies with model-based controllers; develop sim-to-real transfer pipelines; formulate reward functions with safety constraints; maintain training infrastructure; evaluate policy performance; publish research
Seniority
Senior, hands-on IC
