Two featured remote contracts for senior engineers working on AI evaluation and agentic coding. Fully remote, open worldwide.
Browse the roles below, read the full description and apply through the listing. Interviews for roles like these are often run by an AI interviewer, so we have added a short preparation guide and free tools to help you get ready.
Design coding benchmarks and data pipelines used to evaluate frontier AI models. Python expertise and 4+ years of engineering experience required.
Build, evaluate and improve autonomous coding agents. Flexible hours, 5+ years of engineering experience, any major language.
Interviews for roles like these are commonly run by an AI interviewer, so you will likely be talking to software rather than a person. An AI interviewer does not react the way a human does: it takes your answers literally, gives few cues, and follows up on anything vague. Clear, specific, well-paced answers win. Name the concrete thing you built, the numbers, the trade-off you made and what you would do differently.
Senior Software Engineer — AI Evaluation & Benchmarks: For this role, expect questions about designing benchmarks and data pipelines, testing Python code, and judging whether AI-generated code is actually correct, including edge cases and failure modes.
Senior Software Engineer — Agentic Coding: For this role, expect to review code you did not write, to explain how you would break an autonomous coding agent, and to reason aloud through ambiguous, multi-step problems.
Read the invitation instructions carefully and follow them during the interview: the format, timing, tools you may use and any rules on what is allowed.
The best preparation is rehearsing out loud before the real thing. InterviewAce, our interview-prep product, runs mock interviews with an AI interviewer, lets you practise follow-up questions about your own projects and gives you feedback before the real one.
Polish your application with these free tools from InterviewAce.
This is general preparation. We do not run these interviews and cannot say exactly how they will be conducted.