Founding Engineer

About the Role

This is the first full-time engineering hire outside the founding team at an early-stage AI/ML startup that automatically detects failures in agent conversations and opens pull requests to fix them. You will own the entire loop, from ingestion and detection through patch generation and verification, and you will set the engineering standards that everyone who joins after you will follow.

What You'll Do

• Ship end-to-end: take problems from detection through ingest, patch generation, and verification into the product customers use.
• Own the trust surface, including reproduction logic, guardrails, and checks that prevent bad PRs from reaching customer repositories.
• Move from customer conversation to shipped solution independently, without waiting for specs or sign-off.
• Talk to customers directly: read their traces, join their communication channels, and surface what broke before it escalates.
• Make architectural decisions that last: schema design, service boundaries, and tradeoffs between speed and correctness.
• Define engineering culture from scratch, including testing standards, on-call practices, and release criteria.
• Work across the full system: some weeks on detection quality, some on ingest pipelines, some on the review UI that determines whether a customer trusts your PR.

What We're Looking For

• 3 or more years shipping production software with a strong track record of working effectively outside a core specialty.
• Strong TypeScript proficiency across the full stack, frontend through core services.
• Real comfort with data at volume: you think in cost-per-event, not just correctness.
• Experience building and operating AI agents in production environments, with clear opinions on why they fail.
• Experience designing and optimizing high-volume data pipelines with cost awareness.