At a glance
- Employment
- Full-time
- Workplace
- Onsite
- Location
- San Francisco
About the role
- On the RL Data team, you decide what data is used for our agents, including which problems they practice and how they are judged.
- You take weaknesses and turn them into training data to fix them, then show the next model is measurably better.
- You form hypotheses about why data is helping or misleading the model, build the systems to test them, run ablations, and read rollouts.
Responsibilities
- Turning a capability gap into thousands of verifiable tasks
- Building environments that behave like production
- Designing graders for open-ended work that separate good solutions from bad
- Building automated checks that catch reward hacking, leaked answers, broken environments, and invalid paths before data reaches training
- Scaling from a thousand hand-checked tasks to millions with filters and trusted judges
Requirements
- Strong software engineering fundamentals
- Built non-trivial systems in Python such as data pipelines, test or eval infrastructure, sandboxes, or backend services
- Taste for what makes an environment trustworthy
- Careful and systematic approach to understanding aggregate metrics and close reading of raw outputs
- Enjoyment of turning messy, open-ended problems into repeatable, high-quality data production systems
Preferred qualifications
- RL or ML experience is a plus but not required
Skills
- Python
