We need an autonomous, fast-moving engineer with experience building and shipping LLM-powered products and agents who is comfortable working across the full stack and the modern AI toolchain. You should be a power user of Claude Code, Codex, or Cursor, have a track record of taking products from prototype to production, and thrive in a high-ambiguity, high-ownership environment. Bonus points if you have genuine curiosity about government, policy, or regulated industries.
What you'll do:
Build and own the AI agents our team relies on end-to-end, from prototype to production, processing legislation, regulations, and government filings into actionable work product.
Take ownership of retrieval and data pipelines that feed agents accurate, up-to-date legislative and regulatory information from messy, inconsistent government sources.
Work closely with our non-technical policy team on a daily basis, treat them as your product managers and incorporate their constant feedback to make agents smarter and more useful.
Design and maintain eval frameworks to measure and steadily improve the quality and correctness of agent outputs.
Build the product surfaces the team works in every day: drafting tools, monitoring dashboards, and alerting systems.
Shape the company's technical roadmap and help build the technical playbook as we grow, with significant input into what we build and the overall direction of the company.
Create net-new products and improve existing product architecture across the platform.
- Engineering experience in full-stack product work, applied AI/ML, or a closely related field
- Experience at a high-caliber company OR education at a Top 20 school
- Experience building and shipping LLM-powered products and agents (agent workflows, retrieval, evals)
- Power user of Claude Code, Codex, or Cursor
- Ability to collaborate daily and consistently with non-technical policy teams
- Ability to work in-person 5-6 days a week in San Francisco, CA (occasional travel)
- 3-8 years of experience as a fullstack engineer, machine learning engineer, or AI engineer