Software Engineer, Robotics - Isaac Lab

NVIDIA NVIDIA · Semiconductors · Shanghai, China

Software Engineer for NVIDIA's Isaac Lab team, focusing on developing and extending physics simulation APIs for robot learning. The role involves debugging simulation issues, translating research into APIs, and engaging with the robotics community. Requires extensive Python and deep learning stack experience, with a strong background in physics simulation or robotics control, and experience in reinforcement learning and imitation learning.

What you'd actually do

  1. Develop and extend physics simulation APIs for robotics learning, including contacts, joints, sensors, and actuation models.
  2. Debug and resolve simulation instability, non physical behavior, and determinism issues in complex robotic systems.
  3. Serve as a primary technical contact for users and customers, triaging issues from GitHub, external partners and internal teams.
  4. Translate research ideas and physics models into robust, user facing APIs.
  5. Participate in code reviews, system design discussions, and contribute to internal and open-source software.

Skills

Required

  • Python
  • Pytorch
  • Tensorflow
  • Jax
  • rigid body dynamics
  • deformable simulation
  • physics simulation
  • robotics control
  • reinforcement learning
  • imitation learning
  • motion planning
  • trajectory optimization

Nice to have

  • Isaac Sim
  • Isaac Lab
  • Isaac Gym
  • Mujoco
  • supporting users
  • maintaining open source simulation projects
  • Publications in major AI and robotics conferences

What the JD emphasized

  • extensive experience in software development with Python and the deep learning software stack
  • Strong background in rigid body dynamics, deformable simulation, physics simulation, or robotics control.
  • 3 years experience in robotics and simulation workflows, including reinforcement learning, imitation learning, motion planning, and trajectory optimization.

Other signals

  • robot learning platform
  • autonomous systems training
  • physics simulation APIs
  • reinforcement learning
  • imitation learning