Senior / Staff Sim-to-Real Engineer (Tendon-based Robotic Hands)

Job description

Job responsibilities

• Tendon Mechanism Abstraction: Develop and implement simple, computationally efficient, and highly accurate physical abstractions for complex tendon-driven mechanisms in simulation.
• System Identification & Validation: Design and execute empirical testing protocols to identify system parameters and prove that your simulation abstractions accurately reflect physical reality.
• Actuator Characterization: Perform rigorous system ID and control tuning on low friction robotic actuators to ensure high-fidelity torque and effort mapping.
• Rigid Multi-Body Calibration: Drive the kinematic calibration, inertial parameter identification, and effort production verification across our multi-body hand systems.
• Sim-to-Real Pipeline Ownership: Collaborate closely with the RL and Controls teams to ensure the dynamics engine and domain randomization strategies perfectly support zero-shot or few-shot transfer of control policies…
• Technical Leadership: Operate autonomously to identify bottlenecks in the sim-to-real pipeline, propose solutions, and lead the technical execution from concept to deployed hardware.

Job requirements

• Experience Level: 5+ years of industry or development experience post graduation in robotics, controls, or physical simulation, operating at a Senior or Staff level.
• Sim-to-Real Expertise: Proven track record of successfully transferring RL-based control policies from simulation to physical hardware.
• System ID & QDD Actuators: Deep, hands-on experience with control, system identification, and characterization of Quasi-Direct Drive (QDD) actuators.
• Dynamics & Calibration: Strong foundation in rigid multi-body dynamics, kinematic calibration, and inertial identification.
• Hardware Profiling: Proven experience verifying effort production, characterizing friction/backlash, and diagnosing hardware-software discrepancies in complex mechanisms.
• Autonomy: Demonstrated ability to take an open-ended problem (e.g., "our policies fail when the tendon stretches"), design an empirical test, update the simulation model, and resolve the reality gap independently.
• Specific experience modeling, calibrating, or controlling tendon-driven or cable-driven robotic systems.
• Familiarity with modern RL environments and physics simulators (e.g., MuJoCo, Isaac Sim/Gym, Drake, or Bullet).