Robot Learning Engineer
Core
Teach robots to perform complex manipulation tasks from human demonstration, building teleoperation systems and training policies for real-world deployment.
Role type
Senior IC robot learning engineer (manipulation)
Builds
Trained manipulation policies for dexterous robots with custom tooling
Domain
Construction robotics + imitation learning
Deliverable
production ML models
Required skills
imitation learning, PyTorch, Python, ROS2, robot kinematics, Linux, data collection protocols, failure diagnosis
Preferred skills
VLA models, dexterous hands, VR hand tracking, motion retargeting, egocentric video learning, robotics simulation (MuJoCo/Isaac Sim), data vendor management
Technologies
PyTorch, Python, C++, ROS2, Linux, Docker, Git, MuJoCo, Isaac Sim, LeRobot, rosbag, MCAP
Responsibilities
Build teleoperation pipelines using leader arms and VR tracking; Retarget human motion to robot hands in simulation and hardware; Train and fine-tune manipulation policies using imitation learning and VLA models; Define and run evaluation protocols on real robots; Diagnose failures across data, model, and hardware; Supervise external partners for data pipeline infrastructure and labeling
Seniority
Senior, hands-on IC
