Open role
Founding Engineer, Environments and Data
$150,000 to $300,000
The company
A seed-stage startup in San Francisco with an AI agent platform, which also builds environments and benchmarks that frontier AI labs use for training and evaluation. The team is small, in person and led by two founders.
The role
You would turn real agent failures into environments and benchmarks that labs will accept, owning each deliverable from scope to shipping. The role carries a share of the profit on the work you deliver, on top of salary and equity.
What you would do
- Build synthetic environments from scratch, such as a fully simulated company an agent can work inside for weeks
- Own lab deliverables from start to finish
- Mine production agent traces for real failures and rebuild them as benchmarks
- Build graders that hold up against reward hacking
- Design benchmarks whose rankings match how agents perform in production
What we are looking for
- One to five years of engineering or research experience
- Evaluations, benchmarks, RL environments or training datasets built for LLMs or agents
- A background in AI research, AI data, RL environments or agent products
- Comfortable working through raw agent traces to separate real failures from noise
- Suited to an in-person, high-ownership startup
Interview process
- Initial screening
- Technical interview
- Work trial or take-home
Key details
- Salary
- $150,000 to $300,000
- Equity
- 0.25% to 1%
- Experience
- 1 to 5 years
- Location
- San Francisco (On-site)
- Stack
- Python
- Visa
- Visa transfers considered
- Openings
- 2 openings
- Interview
- 3-stage interview process
- Company name
- Shared in confidence once we have spoken
- Reference
- CR-160