What You’ll Be Doing:
Design and implement scalable data generation pipelines using Isaac Lab and Omniverse for learning dexterous, high-DOF robotic tasks.
Extend and optimize NVIDIA’s Isaac Lab/Mimic toolchain for large scale learning from demonstration.
Collaborate with research scientists on cutting-edge problems such as imitation learning, motion retargeting, manipulation, and loco-manipulation.
Integrate large-scale, GPU-accelerated training workflows leveraging NVIDIA’s AI and simulation infrastructure.
Develop modular and reusable environments for benchmark robotic tasks across manipulation, locomotion, and loco-manipulation.
Contribute to open-source initiatives or publish high-quality research and where appropriate.
What We Need to See:
M.S. or Ph.D. in Robotics, Computer Science, Mechanical Engineering, or a related field, or equivalent experience.
4+ years of experience in simulation-based reinforcement learning or robotics software.
Proficient in Python and C++, with experience in GPU programming or CUDA being a strong plus.
Deep understanding of physics simulation tools (Isaac Gym, Mujoco, Bullet, or similar).
Experience with motion imitation, pose tracking, or trajectory optimization using tools like Mimic or RL frameworks (e.g., IsaacRL, RLlib, Stable Baselines).
Proven track record of delivering complex software projects or publishing at top-tier conferences (RSS, CoRL, ICRA, NeurIPS, etc.).
Ways to Stand Out from the Crowd:
Experience developing and deploying sim-to-real policies on physical robot hardware.
Contributions to open-source projects in the robotics, RL, or simulation ecosystems.
Prior experience working with NVIDIA Isaac Gym/Isaac Sim and the Omniverse platform.
Familiarity with large foundation models for vision or language, and how they can interface with embodied agents.
Strong grasp of multi-agent reinforcement learning, curriculum learning, or hierarchical policy design.
משרות נוספות שיכולות לעניין אותך