NVIDIA is looking to build foundation models and full-stack technology for humanoid robots, requiring expertise in physics simulation to develop and optimize these systems.
Requirements
- Proven experience with one or more physics simulators such as MuJoCo, Isaac Sim, PyBullet, Drake, or Gazebo.
- Deep knowledge of state-of-the-art simulation techniques, such as accurate contact dynamics for manipulation and locomotion, and photorealistic rendering for perception.
- Expertise in generating simulation assets, task definitions, and building Gym-style APIs to support neural network training.
- Hands-on experience with deploying and debugging neural network models on robotic hardware;
- Expertise at reinforcement learning and neural network training;
- Contributions to popular open-source simulation frameworks or research publications in top-tier conferences, such as ICRA, IROS, RSS, CoRL.
Responsibilities
- Develop and maintain simulation environments built on frameworks like MuJoCo, and Isaac Lab to support robotics research.
- Implement and test control algorithms and XR teleoperation interfaces for simulated robots.
- Build procedural generation pipelines for diverse environments, object layouts, and robot motions.
- Optimize GPU-based physics simulator performance for large-scale training workloads.
- Import, configure, and validate robot assets in USD format, ensuring successful sim2real transfer.
- Implement Sim2Real pipelines and deploy learned models to physical robots.
Other
- 10+ years of full-time industry experience on robotics and/or physics simulation;
- Demonstrated Tech Lead experience, coordinating a team of robotics engineers and driving projects from conception to deployment;
