The next ten years of AI will not be won in software. They will be won in the physical world. In factories, hospitals, kitchens, fields, and homes. The companies that own that data will own the century.

microagi is building it. We are the data layer for physical AI.

We collect more egocentric human data than anyone in the world. Millions of hours of people doing skilled physical work with their hands. The open research question, and the one this role owns, is what a robot can actually learn from it.

This is a frontier robot learning seat. Dexterous manipulation, learning from human demonstration at a scale nobody else has access to, and the representations that let a skill transfer from a person to a machine.

What You Will Do

• Set the research direction for learning robot policies from human data, and defend the choice.
• Push dexterous manipulation forward. Contact-rich tasks, multi-finger control, and the parts that are still open problems.
• Work out how to turn egocentric human video into something a robot can learn from. Action representation, the embodiment gap, retargeting, and what actually survives the transfer.
• Train and evaluate policies in simulation and on real robots, and be honest about which results hold outside the lab.
• Publish. We want the work in the open and we want you contributing to the state of the art, not sitting on it.
• Work closely with the pretraining, reinforcement learning, and deployment people rather than in parallel to them.

Requirements

• A track record at the frontier of robot learning. Multiple publications at top venues, or work of the same calibre that shipped instead of being published.
• Depth in dexterous manipulation, imitation learning, vision-language-action models, or diffusion policies.
• Experience learning from human demonstration, egocentric video, or teleoperation data.
• Strong PyTorch or JAX. You write the training code yourself.
• Experience running experiments on real robots, not only in simulation. Hands-on with Isaac Sim, MuJoCo, ManiSkill, or equivalent.
• A PhD is the common route into this seat but it is not a requirement. We care what you have done, not where you did it.
• You can hold a direction for months and still kill it when the evidence says to.
• High agency - you don't wait to be told what to do.
• Fluent in English.

Nice to Have

• Experience with whole-body control, teleoperation rigs, or dexterous hands.
• Experience with egocentric datasets such as Ego4D, Ego-Exo4D, or EPIC-Kitchens.
• Open-source work that other researchers build on.

#LI-DNI