Building the foundational models for the next era of human and machine intelligence.
Research Scientist
You’ll ship research that redefines how models reason, adapt, and teach—not in a demo, but in systems people rely on daily. In your first year, expect to lead a core thread of investigation from hypothesis through working implementation, collaborating with a handful of engineers and domain experts who treat rigor as a baseline, not a goal. We publish openly, build in the open, and measure progress by what our models can actually do under real constraints. The work sits at the intersection of deep learning fundamentals and applied education, where small gains in reasoning compound into outsized human outcomes.
What they're looking for
- 4+ years of hands-on research experience designing, training, and evaluating large-scale models in PyTorch, with a portfolio of published or shipped work to show for it
- Fluent across the stack from experiment scripting in Python to lightweight front-end tooling in TypeScript when research demands it
- Track record of pursuing ambitious, open-ended questions and converting partial signals into clear next experiments
- Comfortable working on-site in San Francisco or New York with a small, high-trust team where your voice shapes research direction from day one