Helix AI Engineer, Reinforcement Learning

What the role actually is

Figure is hiring a reinforcement learning engineer for Helix, its core AI system for humanoid autonomy. The role focuses on policies that learn from interaction, feedback, logged robot data, and simulated experience.

The listing is centered on robot learning infrastructure and evaluation, including reward modeling, exploration, policy robustness, and long-horizon behavior in embodied systems.

What you would work on

• Design and implement reinforcement learning algorithms for embodied agents
• Train policies from interaction, feedback, and large-scale robot experience
• Build reward modeling, exploration, and credit-assignment methods
• Create evaluation frameworks for policy stability and generalization

What they are asking for

• Experience applying reinforcement learning in complex environments
• Ability to work across online and offline RL settings
• Comfort building scalable training systems, rollouts, and experiment workflows
• Interest in integrating RL with pretraining, video, generative, and robot learning teams

Why this one is worth a look

Helix is Figure's public bet on general humanoid intelligence. This role is not a product design job, but it belongs on the board because learned robot behavior is becoming one of the central user experience questions for humanoids.