InLoop builds warehouse robots that run fully autonomously in customer warehouses today. The fleet is growing quickly (20+ robots), and the demand is huge.
We're hiring a research engineer to make our robots learn precise, contact rich tasks in simulation and carry them over to real hardware. You will be working on the real customer use-cases, and getting the best feedback, directly in a field.
This is research that ships. What you train runs on real robots in weeks, not years.
**What you'll work on**
* **Manipulation with RL.** Train reinforcement learning controllers for precise manipulation tasks.
* **Sim to real.** Get policies trained in simulation working on our real arms, and close the gap when they don't.
* **Control.** Combine learned policies with classical and learned control where it makes the robot more precise or safer.
* **Into the stack.** Bring what works into our autonomy stack so it runs on the fleet.
**What we're looking for**
* Hands on RL for robot control, trained in simulation and run on real hardware
* Large scale simulation (Isaac Sim, Isaac Lab, MuJoCo or similar)
* Strong control fundamentals
* A track record of shipping: code on real robots, papers or open source
**Nice to have**
* Experience with Vision language action models or imitation learning
* Experience with Code-as-a-Policy
* PhD with a track record of publications (CVPR, ICRA, ICLR, CORL, and etc.)
**How to apply**
Apply here and include a link to a policy you trained running on a real robot, a paper or a repo. That tells us more than a CV.
Visa sponsorship is offered for this role.
