Reinforcement Learning Engineer - Manipulation
Core
Build highly performant and robust manipulation policies for humanoid robots using reinforcement learning in both simulation and physical reality.
Role type
Senior IC reinforcement learning engineer (robotics manipulation)
Builds
Language-vision conditioned manipulation policies for HMND-01 Alpha humanoid platform
Domain
Robotics + Deep Learning (Reinforcement Learning)
Deliverable
production ML models
Required skills
Reinforcement learning with deep neural networks, LLMs/VLMs/image/video generative models, Python, PyTorch/JAX, simulation-based task construction, behavior cloning, real-world RL training pipelines
Preferred skills
Robotics simulators (Isaac Sim, MuJoCo), large-scale RL infrastructure (Ray), OpenVLA/Physical Intelligence frameworks, top-tier conference publications
Responsibilities
Train language-vision conditioned manipulation policies via RL in simulation and real world, construct diverse manipulation tasks in simulation, partner with teleoperations for trajectory collection, establish real-world RL training pipelines, experiment with sim-to-real transfer
Seniority
Senior, hands-on IC
