Simulation Engineer - Design and Behavior - Dexterity

Apptronik Apptronik · Robotics · HQ · Software Engineering

Apptronik is seeking a Simulation Engineer to own the "Design-to-Simulation" loop for their next-generation dexterous hands. This role involves ingesting hand designs into simulation, constructing high-fidelity interaction scenes, implementing teleoperation pipelines, and using simulation to train and test advanced manipulation behaviors, collaborating with controls teams to prototype skills. The engineer will also create automated analysis tools for workspace and kinematic analysis, and partner with the hardware team to measure and reduce the sim-to-real gap.

What you'd actually do

  1. Develop tooling to rapidly convert CAD into Universal Scene Description (USD) assets with accurate collision meshes, joint limits, and kinematic parameters optimized for Isaac Sim.
  2. Design and implement the evaluation environments in Isaac Lab, including the multiple dexterity benchmarks and scenarios. You will own the modeling of the physical properties for high-fidelity photorealistic dynamic simulation.
  3. Implement and maintain the interface between human input devices (VR gloves, controllers) and the simulated hand. You will ensure the teleoperation stack performs identically on simulated hands as it does on real hardware.
  4. Create automated analysis tools via Python scripting in Isaac Sim to visualize reachable workspace and manipulability ellipsoids for new hand designs immediately upon import.
  5. Partner with the hardware team to iteratively measure and reduce the sim-to-real gap. You will run side-by-side comparisons of grasping stability to tune simulation parameters until they match reality.

Skills

Required

  • Deep experience with modern robotics simulators, like Isaac Sim or MuJoCo.
  • Strong understanding of the Universal Scene Description (USD) file format, including composition arcs, layers, and stages.
  • Strong Python skills for scripting, pipeline automation, and data analysis.
  • Solid understanding of kinematics, rigid body dynamics, collision geometry generation, and spatial math.

Nice to have

  • Proficiency in C++ is a plus.

What the JD emphasized

  • train and test advanced manipulation behaviors

Other signals

  • robotics simulation
  • dexterous hands
  • train and test advanced manipulation behaviors
  • sim-to-real validation