The role
- COMP
- $150K - $240K
- EQUITY
- <p>Meaningful early-stage equity grant.<p>
- LOCATION
- San Francisco
- WORKPLACE
- On-site
- EXPERIENCE
- 0 - 5 years
- VISA
- None, Visa transfers
- STACK
- Python
- INDUSTRY
- AI, Data
/roles — ROLE_288
Research Product Manager (RPM)
Small, recently founded AI data lab producing training environments and datasets for frontier model developers
The company
Early-stage AI data lab that builds simulation environments and generates synthetic data so leading AI labs can train more capable agents with reinforcement learning.
- STAGE
- seed-stage
- FUNDING
- seed funding
- TEAM
- ~10 people
- FOUNDED
- 2024
- BACKING
- VC-backed
JD — the work
About the role
This is a technical product role that sits between research and delivery at a small, recently founded lab producing training and evaluation data for leading AI labs. You would find where frontier models fall short, design the experiments and datasets that close those gaps, and run each engagement from scoping to weekly data handoffs. The company's view is that data, not compute, is now the main bottleneck for agents, robotics and multimodal models, which puts this work close to the frontier.
What you'll do
- Pinpoint where frontier models underperform and set improvement targets together with researchers at partner labs
- Run headroom analyses, benchmarks and experiments on internal simulation tooling to lift model performance in target verticals
- Turn loosely defined requests from partner labs into concrete research questions, then rally internal teams to tight deadlines
- Set up data curation, either by sourcing outside domain experts or by building synthetic generation with research engineers
- Read the latest research closely and apply what matters to how datasets and evaluations are designed
- Own each project from scoping to weekly dataset shipments, iterating on customer feedback in one to three month cycles
- Link researchers and operations staff, explaining results plainly and turning customer problems into evals and datasets
What they're looking for
- 0 to 5 years of experience in technical product management or ML research
- Deep technical and research instincts, plus comfort with ambiguity at a fast-moving early-stage startup
- Fluency with coding agents, benchmarks, and how LLMs are evaluated
- Proven ability to deliver technical projects from start to finish
Nice to have
- Experience founding a company of your own
- Time at an AI data company or a frontier AI lab