Description
The role owns dynamic-model accuracy for a humanoid robot, including system identification on 4NE1 hardware, MuJoCo and Isaac Sim model authoring, real-time floating-base EKF/UKF state estimation, sim-to-real calibration, and debugging model-driven control and estimation failures. It requires an MSc or PhD in a relevant engineering field, at least four years of real-hardware state estimation or system identification experience, and proficiency in C++ and Python, with humanoid, differentiable simulators, Pinocchio, MuJoCo, and factor-graph estimation experience preferred.
