We are looking for a Robotics ML Engineer with 0-7 years of experience to own robot learning end to end, from data to deployed policy. We build our own hardware and train our own models to automate electronics manufacturing. We just passed 100 hours of robot data and are scaling collection fast. This is applied research: you will iterate on real problems on real robots, not research for its own sake.
What will you be doing?
Build pipelines that ingest and quality-check terabytes of robot data a day
Post-train models and set up fast, reliable training pipelines
Calibrate hardware and run data QC so models perform well on our setup
Set the company's standards for data processing and model training
Keep up with new robot-learning research and apply what works
This is a direct-hire, on-site role in San Francisco, CA.
- Hands-on experience working with physical robotic systems (robots, manipulators, drones, or autonomous vehicles)
- Strong Python proficiency
- Practical proficiency with PyTorch, JAX, or an equivalent framework
- Willingness to work on defense and government robotics applications
- 0-7 years of experience as a robotics machine learning engineer
- Applied machine learning engineering experience rather than pure research
- Trained and deployed learned policies on a real robot that moves (manipulation, driving, drones)
- Robot learning depth, such as post-training policies and training pipelines
- Experience at target robotics or autonomous vehicle companies (Physical Intelligence, Skild AI, Dyna Robotics, Dexterity, Ambi Robotics, Zipline, Waymo, Cruise, Skydio, Shield AI, Tesla, Anduril, SpaceX, Generalist, Walden, REI)
- Defense-tech experience
- C++ or Rust, low-level GPU work, distributed training or eval infrastructure
- Strong engineering university background (Stanford, Berkeley, MIT, CMU, Georgia Tech or Columbia) preferred