About the role- You'll work on whatever is currently the hardest problem in the software that carries our models from training to production: data infrastructure, services, pipelines, APIs, deployment, and the internal platform our researchers and operators depend on.
- Scope is set by what's blocking, not by a team boundary. Expect to move between problems.
- You'll know your first project before you accept an offer.
- Your work reaches a real robot doing a real job in days or weeks, not quarters.
What we look for- You believe physical intelligence is the most important unsolved problem in AI, and that it gets solved by shipping into the messy real world rather than by waiting for a cleaner simulator.
- You are an exceptional engineer by any measure of software engineering.
- You have a high tolerance for ambiguity and the resilience to keep making progress when the problem, the requirements, or the data change underneath you.
- You take pride in your work and want to be measured on it.
Prior robotics experience is welcome but not required. Several of our strongest engineers came from distributed systems, compilers, and large-scale backend or training infrastructure.
