Build with us
Come build with us.
We’re looking for people who want to make robots learn — and who’d rather try it than talk about it.
Who we’re looking for
People who are curious about robot learning and want their hands on it: reinforcement learning, world models, control, perception, and the unglamorous parts in between. You don’t need a PhD or a publication record. You need to be the kind of person who reads a paper and immediately wants to run it on something real.
What helps
- Python and PyTorch, at the level where you can read someone else’s training loop
- Some RL — even if it’s only gym environments you broke and fixed yourself
- Simulators: MuJoCo, Isaac, Genesis, whatever you’ve actually used
- Or no ML at all, and instead real hardware experience: motors, sensors, things that vibrate loose
What matters more
That you finish things, and that you can say clearly what broke. We’re early enough that taste and persistence count for more than credentials.
How to reach us
Tell us what you’re working on — a repo, a video of something twitching on your desk, a half-finished idea. That’s a better introduction than a CV.