Research Scientist

We're building robots that learn through exploration in the real world. We're looking for research scientists with strong foundations in reinforcement learning, multimodal representation learning, or large-scale model training.

Qualifications:

• You’ve worked on one of:
  • Self-supervised, goal-conditioned, or unsupervised RL
  • Large-scale pretraining (language, multimodal, image, or video)
  • RL post-training, reasoning, or tool use
  • Robotics models, especially methods that involve large-scale pretraining
  • Learning methods for low-level motor control
• You’ve worked on large GPU clusters, and are comfortable working with Kubernetes
• You believe that research and engineering are two sides of the same coin
• You strive to find simple, expressive metrics, and to measure them accurately
• You value scientific integrity, and seek to understand the true effect of different interventions

Nice to have:

• Experience with JAX
• A strong proof-based math background, for example:
  • Measure-theoretic probability theory
  • Stochastic processes
  • Optimization theory

We care much more about what you can do than any specific credential. We're interested in published work or lab experience, but equally in strong open-source contributions or personal projects.

We're a small fast-moving team working together in person, in San Francisco. If you're excited about getting robots to learn from the real world, we'd love to talk. To apply: send an email to hiring@pantograph.com with "Research Scientist" in the subject line, and include a CV and a description of your research interests or what excites you about this area.