Lemma is production monitoring for AI agents. We catch the silent failures your observability tools and evals miss (think bad tool calls, lost context, and infinite loops) before your users find them. ### Why this role exists Agents break silently. They call the wrong tool, forget what the user said three turns ago, and loop until someone pulls the plug. Soon they'll be responsible for the majority of the world’s economic work, and most teams won't even know when they fail.
Making agents reliable is the problem Lemma exists to solve. That means catching the unknown unknowns, the failures nobody thought to write an eval for, and closing the loop end to end so they get fixed, not just flagged. It's the foundation for building agents people can actually trust and the future of self-improving systems. Finding failures is only half of it. The other half is a human problem. Every issue we surface is more homework for a team that already has too many priorities.
If we can't decide what matters, rank it, and put it in front of the right person at the right moment, we've just built a noisier inbox. This role owns that half: how issues get surfaced, prioritized, and acted on; how Lemma fits into the workflows teams already live in; and how humans and agents work together to close the loop. ### What you'll do * Turn thousands of signals into the three things a team should fix today * Own the loop end to end: detection, prioritization, alerting, triage, and the fix * Integrate Lemma into the tools teams already use (Slack, Linear, GitHub, coding agents) so acting on an issue takes less work, not more * Shape how humans and agents share the work of fixing things * Ship across the full stack in TypeScript, from the core service to the product people look at every day * Work directly with customers, and set the engineering standards everyone after you will inherit ### What we're looking for * High slope over years of experience.
New grads and dropouts welcome * A track record of shipping things people actually use * Strong TypeScript, frontend through backend * Real product taste. Showing less is usually harder, and better * Bonus: you were the customer once. You ran agents in production and got burned ### Who you’ll work with You'll join a team of dropout founders and engineers from Amazon, Together, and Zoom. We've been founding operators at unicorns and at startups that went on to be acquired.
Onsite in San Francisco. We don’t sponsor visas.
Job details are sourced from the employer's original posting.
Open job posting