Research Scientist Intern, Robotics Foundation Models Post-Training (PhD)
Meta
Zurich, Switzerland · Robot Learning & AI
Develop scalable offline reinforcement learning methods to post-train robot world action and vision-language-action models using large, mixed-quality datasets. Work includes reward learning, policy optimisation and studying how performance scales with data and compute. Intended for PhD candidates on a 12 to 24 week internship.
View on in-house- Company
- Meta
- Location
- Zurich, Switzerland
- Level
- Intern
- Specialism
- Robot Learning & AI
- Posted
- 20 October 2026