Skip to main content

Figure

Helix AI Engineer, Robot Learning

San Jose, CAmidAdded 1 month ago

About this role

Figure AI is seeking a Robot Learning engineer to develop and deploy visuomotor manipulation policies for humanoid robots. You'll work across the full pipeline from real-robot data collection to model training and deployment, focusing on practical real-world performance for tasks like grasping, pick-and-place, and assembly.

What you'll do

  • Design, train, and deploy learning-based visuomotor policies for humanoid robot manipulation tasks
  • Develop manipulation behaviors including grasping, pick-and-place, object reorientation, and bimanual manipulation
  • Apply behavior cloning, reinforcement learning, and VLA techniques to robot control problems
  • Manage the complete pipeline from data collection on real robots to model evaluation and deployment
  • Collaborate with perception, controls, systems, and hardware teams to integrate policies into the autonomy stack
  • Write robust, production-quality software that runs reliably on physical robots

What they're looking for

  • Robot manipulation and visuomotor control
  • Behavior cloning and reinforcement learning
  • Python and/or C++ for robotics and ML
  • Deep learning frameworks (PyTorch)
  • Real-world robotic system experimentation and debugging
  • Classical vs. learning-based robotics tradeoff analysis
  • Hands-on robot deployment experience
  • Software testing and production system reliability
Apply with Autofill

Opens the application — the Jobs AI extension fills it for you. Set up autofill

Opens the official application on the employer’s site. No login required.

Figure

Figure develops advanced humanoid robots powered by AI technology. The company is hiring engineers across mechanical design, firmware development, manufacturing, quality assurance, and security to build and refine its autonomous robotic systems.

View all jobs at Figure

Likely interview questions

  • Walk us through a specific robot learning project where you deployed a visuomotor policy on real hardware. What were the main challenges in the real-to-sim or sim-to-real transfer, and how did you address them?
  • Describe your experience with behavior cloning and/or reinforcement learning for manipulation. Which approach have you found more effective for different types of tasks, and why?
Unlock all 7 questions free — and practice them live →