About the role
About the Company
RL environment startup, YC-backed, ~20 people · San Francisco · On-site, 5 days · $200,000 – $300,000 + 0.15–0.3% equity. AI agents are only as good as the places they learn. Our client builds the sandbox environments and evaluation benchmarks that frontier labs use to train computer-use and browser agents — simulated applications, verifiable rewards, and the infrastructure to run thousands of them reliably. They've narrowed onto financial services: investment banking, private equity, consulting.
About the Role
You'd be one of the first platform engineers, reporting to the CTO, with a mandate to make environment creation dramatically faster. The tooling underneath: orchestration, isolation, reproducibility, and the analytics that show where the time and money actually go.
Responsibilities
- Rebuild how environments are provisioned and torn down, so a new one takes hours rather than days
- Build the cost and bottleneck analytics the team currently reconstructs by hand
- Work on domain-specific verifiers with the people writing the tasks
Qualifications
You’ll fit if you’ve run containerised or sandboxed workloads at scale (Kubernetes, microVMs, Firecracker), written production Go, Python or TypeScript, and worked somewhere young enough that you set the patterns rather than inherited them.
Preferred Skills
Five days a week in San Francisco. Deliberate, and not flexible.
Pay range and compensation package
$200,000 – $300,000 + 0.15–0.3% equity
Equal Opportunity Statement
Our client is committed to diversity and inclusivity.
Requirements
Nice to have
- Five days a week in San Francisco. Deliberate, and not flexible.
- Pay range and compensation package
- $200,000 – $300,000 + 0.15–0.3% equity
- Equal Opportunity Statement
- Our client is committed to diversity and inclusivity.