Research Scientist, Reinforcement Learning - Atlas
Core
Design, train, and deploy reinforcement learning policies for whole-body mobile manipulation and dexterous tasks on humanoid robots in unstructured environments.
Role type
Research Scientist, Reinforcement Learning (Robotics)
Builds
RL policies for Atlas humanoid robots
Domain
Robotics, Reinforcement Learning
Deliverable
production ML models
Required skills
Reinforcement Learning, Python, C++, Robotics fundamentals (kinematics, dynamics), ML frameworks (PyTorch, TensorFlow, RLlib), Simulation environments (Isaac Sim, MuJoCo)
Preferred skills
Physical robot deployment, Locomotion, Bimanual manipulation, Whole-body control, Open-source contributions, Top-tier conference publications
Technologies
Isaac Sim, MuJoCo, PyTorch, TensorFlow, RLlib
Responsibilities
Design and train RL algorithms for mobile and bimanual manipulation; Develop production-ready Python and C++ code; Build and leverage high-fidelity simulation environments; Integrate policies with control and software stacks; Deploy and debug policies on real hardware; Participate in design reviews and experimental planning
Seniority
Senior, hands-on IC
