Reinforcement Learning Engineer, Policy, Optimus

Tesla Tesla · Auto · Palo Alto, CA · Tesla AI

Develop and deploy end-to-end robotic learning systems (reinforcement/imitation learning) for humanoid robots, focusing on complex physical tasks, locomotion, manipulation, and language-conditioned tasks from vision. Ship production-quality, safety-critical software utilized by thousands of robots.

What you'd actually do

  1. Develop end-to-end robotic learning with either reinforcement or imitation learning
  2. Reinforcing correct set of actions, rewarding correct behavior and negating incorrect behavior (with real-time action/reward feedback loops)
  3. Perform a large number of instructions and generalize new tasks with different objects and environments
  4. Learn to perform dexterous tasks using high degree of freedom hands
  5. Learn different robot policies to solve language-conditioned tasks from vision
  6. Ship production quality, safety-critical software

Skills

Required

  • Experience in end-to-end robotic learning, with either imitation or reinforcement learning
  • Experience writing production-level Python (including Numpy and Pytorch)
  • Experience with distributed deep learning systems
  • Proven track record of training and deploying real world neural networks

Nice to have

  • Exposure to robot learning through tactile and/or vision-based sensors

What the JD emphasized

  • end-to-end robotic learning
  • reinforcement learning
  • imitation learning
  • production quality, safety-critical software
  • real world applications

Other signals

  • end-to-end robotic learning
  • reinforcement learning
  • imitation learning
  • production quality, safety-critical software
  • deployed to and utilized by thousands of humanoid robots