About the role
Responsibilities:
- Own the end-to-end locomotion stack, interfacing high-level VL A semantic goal points (5-20 Hz) with low-level deterministic motor controllers (500 Hz).
- Develop and implement nonlinear Model Predictive Control (MPC), Whole-Body Control (WBC), and Deep Reinforcement Learning locomotion policies.
- Oversee hardware validation of gait transitions, obstacle clearing, push recovery, and stair traversal.
- Construct massively parallel Sim-to-Real (Sim2Real) training pipelines within NVIDIA Isaac Lab / Isaac Sim to produce robust gait policies across varying terrain geometries.
- Implement deterministic state estimation, contact detection, and slip-recovery algorithms operating at 500 Hz.
- Integrate high-speed motor communication protocols over CAN-FD and EtherCAT networks.
Requirements:
- Master's or Ph. D. in Mechanical Engineering, Robotics, or Mechatronics.
- Minimum 8 years of practical experience developing dynamic legged locomotion, multi-body kinematics, and real-time control.
- Expertise in modern C++, Pinocchio, Drake, OCS2, Isaac Gym, and PREEMPT_RT Linux environments, C++20, Quadratic Programming (QP), and Kinematics and Dynamics.