Job description
An AI robotics company building synthetic data and robot-learning models is hiring its first robot-learning engineer in the U.S. and a founding member of the U.S. team. You will train robot policies on synthetic data, starting with manipulation and grasping, establish baselines on public eval suites, retrain under the same protocols, and report what moved, including failures. You start as an individual contributor and then hire and lead the Bay Area robotics team.
What you'll do
- Train robot policies (VLA fine-tuning, imitation learning / diffusion policy, or RL in simulation)
- Establish baselines on public eval suites and run controlled comparisons with fixed seeds and protocols
- Turn results into written data specs for the data team
- Publish and demo (CoRL, RSS, ICRA) and join customer technical conversations
- Grow into hiring and leading the Bay Area robotics team
Compensation
- $260,000 to $300,000 base, plus competitive equity
Location
San Francisco, three days a week in office.
Requirements
- 6+ years hands-on robot learning (industry or university lab)
- Trained and evaluated robot policies in simulation (Isaac Sim/Lab, MuJoCo, or equivalent)
- MS or PhD in robotics, ML, or a related field
- Policy-learning methods: VLA fine-tuning (OpenVLA, pi0, GR00T), imitation learning (diffusion policy, ACT), or RL in sim
- U.S. work authorization; this role does not offer visa sponsorship
Nice to have: manipulation/grasping or navigation policy training.
Tech stack: Isaac Sim/Lab, MuJoCo, Python, PyTorch, diffusion policy, VLA models, sim-to-real.