Software Engineer - Human Motion Data
Core
Architecting robust motion data pipelines to translate raw human movement into actionable, dynamically feasible data for whole-body reinforcement learning in humanoid robots.
Role type
Software Engineer - Human Motion Data
Builds
Motion data pipelines integrating mocap, teleoperation, and generative AI models for Apollo humanoid robots
Domain
Robotics / Embodied AI / Reinforcement Learning
Deliverable
production ML models
Required skills
Robot kinematics (FK/IK), rigid body dynamics, Python scripting, motion capture data processing, retargeting libraries (GMR, Omni-retarget), game engines (Unreal/Unity), physics simulators
Preferred skills
C++, hardware experience with physical robotic platforms, sim-to-real pipelines, portfolio of robotics projects
Responsibilities
Design and maintain end-to-end motion data pipelines; Implement kinematic and dynamic retargeting pipelines; Develop tools to process and clean raw human demonstration data; Build simulated environments for procedural motion generation; Generate high-volume trajectory datasets for RL training; Write automated pipelines for data flow between sources and RL infrastructure
Seniority
Mid-level, hands-on IC
