Researcher, Robot Learning

Highlights

• Own generalist policies end-to-end from dataset to live humanoid robot.
• Build self-improvement loop so Asimov gets better the more it works.
• Ship open-source learning systems turning data into general-purpose behavior.

Ideal candidate

Hands-on robot learning researcher who lives for real hardware and data-driven iteration. You want your work running on humanoids in the open, not just in papers, and obsess over how robots learn new skills on their own.

Requirements

• Deep hands-on experience training robot policies via imitation learning and/or reinforcement learning
• Experience taking learned policies from research code to physical robots
• Fluent in Python and PyTorch with multi-GPU cluster training experience
• Strong understanding of dataset composition, quality and diversity for policy performance
• PhD or equivalent in robot learning, ML or related field, or equivalent portfolio
• Clear communication via papers, blogs, talks or well-documented code
• Uses AI as a real collaborator in research