Careers · Engineering
Founding Engineer, Environments + Data
Build the environments that tell us when agents can do a company's work, and when they pass human level.
San Francisco, on-site five days a weekFull-timeVisa sponsorship$150K to $300K base, plus equity and bonuses
The role.
Public benchmarks miss most of what companies need agents to do. We see that work every day because our agents do it.
You'll turn it into synthetic environments that predict how agents do in production, with every email, record and file generated from scratch.
The other engineering track is Founding Engineer, Agents. Apply to whichever fits, or both.
What you'll work on.
- 01A simulated company an agent can work inside for weeks.
- 02Graders that hold up when the agent knows more than any human reviewer.
- 03Benchmarks whose rankings match production.
What we look for.
- 01You've built environments or benchmarks a frontier lab trained or evaluated on.
- 02You've found a real flaw in a widely used benchmark.
- 03You've caught a model reward hacking your grader, and fixed the grader.
- 04You can simulate software like an inbox or a CRM well enough to survive an agent's tool calls.
- 05You're kind and easy to work with, especially when things get hard.
- 06You think agents will beat people at most computer work within a few years.
Bonus points.
- 01You've started a company, or been part of YC, SPC, Thiel or similar.
- 02You've competed at a high level in math, programming, debate or games.
What we pay.
- 01Base$150K to $300K
- 02EquityMeaningful founding equity
- 03BonusesDetails when you apply
We're flexible on the split between base and equity.
How to apply.
Email founders@doe.so with the role in the subject line, a link to the best thing you've built and a few lines on why it was hard.