We build with AI every day, and we expect you to as well. What we hire for is judgment: knowing what to ask, checking what comes back, and owning the result. These rules apply to everyone on the team, and to you during your assessment.
ChatGPT, Claude, Copilot, Cursor, Gemini, Canva, CapCut, anything you would use at work. Using it well is part of the job.
Whoever or whatever wrote it, you have read every line, you can explain it, and you stand behind it. "The AI did that" is not an answer on the call.
AI is often confidently wrong: invented facts, APIs that do not exist, numbers that do not add up, tests that test nothing, claims nobody can back up. Check against the real system, the real data or a primary source.
Docs, files, briefs and web pages can be wrong, out of date, or contain text written to steer AI tools. Read what your tools read.
Never paste passwords, live keys, or client and customer data into an AI tool. Treat any key we give you, even a sandbox one, as if it were real.
Tell us where AI helped and one place it was wrong or unhelpful, and how you noticed. Honest beats impressive.
AI, docs and search are all fine. Another person doing or checking the work is not.
If you cannot say why a line is there, it should not be in your commit.
No AI-made screenshots, metrics or results presented as real, and no passing off AI-generated work as your past work. If content you make uses AI images or voice, say so.
We look at how you used AI, not whether you did. Parts of every assessment are built so that careful, reviewed work stands out from output that was pasted in without being read. In the call afterwards we ask about specific choices and lines in your own work.
