About the job
Isaac Lab is NVIDIA’s flagship product for robotic learning, supporting both large-scale parallelized reinforcement learning and teleoperation and imitation learning. We maintain an open-source project for Isaac Lab that has been acknowledged by the robotics community and has been used by top research labs and the industry as a core simulation tool for training robots. Our team is seeking a software engineering intern to join the Isaac Lab team and propel our flagship platform for robot learning to new heights! Our mission is to become the industry's leading tool, redefining how autonomous systems are trained and crafting the future of robotics and AI. Come join the team and see how we can make a lasting impact on the world!
Responsibilities
Develop the next features for our platform, such as integration with Newton simulation, scalable perception-in-the-loop reinforcement learning and learning from demonstration via tele-operation.
Participate in open-source development of Isaac Lab, including engaging with the robotics industrial and research communities.
Scale training massively in the cloud, while ensuring the highest performance with extensive benchmarking, profiling, and optimizations.
Build and improve the robot asset pipeline, importing URDF/MJCF models, authoring new environments, and validating physical fidelity against real-world hardware.
Bridge the sim-to-real gap by working on domain randomization, system identification, and transfer techniques that move policies from simulation onto physical robots.
Qualifications
Minimum
Pursuing MS or PhD degree in Computer Science, Mechanical Engineering, Robotics, or similar program area.
Experience in software development with Python and the deep-learning software stack (Pytorch, Tensorflow, Jax, etc.).
Experience with robotics and simulation workflows, including reinforcement learning and imitation learning in simulators such as Isaac Lab, Isaac Gym, MuJoCo, or other physics-based simulators.
Preferred
Experience training a robot in simulation and deployed the policy sim-to-real.
Publications in major AI and robotics conferences.