Software Engineer, Platform
San Francisco, Full-Time
Platform software engineers experiment with the latest models and build the infrastructure and tooling that powers our data creation engine.
Sample things you might work on:
Build background agents to scrape and process 300K+ public PRs merged everyday on Github into usable data for model training
Sprint for two weeks to build an agent that will automate a meaningful chunk of our work at scale
Build fault-tolerant infrastructure to run coding agents for 12+ hours
Create new SDKs that let us create new coding datasets with a few lines of configuration
You might be a good fit if:
You're a strong generalist who can building anything from 0 to 1
You've been early at a startup and grown with it
You have good taste for experimenting with LLMs to build automations
You've built systems that run at production scale
About
Our mission is to build system that let coding agents create their own training data autonomously. We're a data research lab solving hard problems in both research and engineering—figuring out what training data works, then building systems that generate it at massive scale.
Our founding team is small and talent dense, we were early at companies like Cursor, Prime Intellect and Browserbase. We care about ownership, solving hard problems, and being the best at what we do.
We're in-person in North Beach, San Francisco.
Interview Process
We usually have 2-3 short technical interviews before an onsite where you work on a small project with our team.
Ready to apply?