Helix AI Engineer, Robot Learning

What the role actually is

Figure is hiring a Helix AI engineer focused on robot learning for humanoid manipulation. The listing emphasizes visuomotor policies and real-robot deployment rather than only offline model development.

The role sits on Figure's push toward general-purpose humanoid robots, so the technical work is tied directly to whether manipulation behavior works on hardware.

What you would work on

• Design, train, evaluate, and deploy learning-based visuomotor policies
• Develop manipulation capabilities for humanoid robots
• Run experiments that connect model behavior to real robot performance
• Work with the broader Helix AI effort on embodied intelligence

What they are asking for

• Strong robotics learning background with practical policy training experience
• Experience deploying learning systems on real robots or similarly constrained hardware
• Ability to evaluate robot behavior beyond benchmark metrics
• Willingness to work onsite from Figure's San Jose office

Why this one is worth a look

Humanoid companies live or die by whether their policies survive contact with the real world. This role is worth tracking because it is explicitly about deployed visuomotor manipulation, not just training models in a lab setting.