You'll own how the agent checks its work: the checks it runs, the evaluations we test it against, and the tools that tell us when an answer is wrong. You'll work directly with Theo and ship to customers every week.

What you’ll do

Build and extend the checks the agent runs before it answers: joins, dates, missing values, and results that don’t look like the data
Turn real customer questions, shared with permission, into evaluations that run on every change
Trace wrong answers to their cause and fix the whole class of error, not just the one case
Make the agent’s reasoning readable to the people who use it

What you’ll bring

Several years of writing production software. We use Python and TypeScript.
Strong SQL and a feel for how analytics goes wrong
Experience shipping products built on language models, or a clear reason you want to
Care about the person reading the answer

Nice to have

Time on a data team
Evaluation tooling you’ve built before

How to apply

Send a short note and something you've built to careers@logithq.com.

Apply by email

Logit is an equal opportunity employer. We don’t discriminate based on race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, veteran status, or any other protected status.

Pursuant to the San Francisco Fair Chance Ordinance, we will consider for employment qualified applicants with arrest and conviction records.