Job Description

Are you the kind of engineer who enjoys building complex systems that help models learn, not by training them directly, but by shaping the worlds they inhabit?

This team builds large-scale environments and benchmarks that frontier AI labs use to test and steer their models. Their goal is to make reinforcement learning measurable, creating rich, hyperrealistic simulations where agents can reason, act, and be safely evaluated.

You’ll work at the intersection of software engineering, reinforcement learning, and experimental research, designing the frameworks and pipelines that let agentic AI systems act, learn, and improve through interaction, not static data.

You'll Bring

• Strong Python and software fundamentals who enjoy building ML infrastructure.

• Experience in reinforcement learning, rewards, environment dynamics, evaluation loops.

• Worked with browser/API simulations (Playwright, Selenium) or distributed compute.

• Experience with open-ended problem spaces and a desire to shape the tools driving safe AGI progress.

It’s a technically deep team of ML engineers and researchers from leading labs and tech companies, developing the simulation and evaluation backbone for next-generation agents.

Compensation: $200,000–$250,000 base + equity
Location: San Francisco (on-site, relocation supported)

All applicants will receive a response.