Your interview, without the mystery.

A human decides. The AI only suggests.

The AI grades one thing: whether your code shows the technical skills the job asks for. It produces a score and a written reason. A real person reads that, weighs it, and makes every hiring decision. The AI cannot reject you. No exceptions.

What we watch, and what we never record

During the assessment, we take note when:

• you switch tabs or leave fullscreen
• a large block of code is pasted in at once
• a second display or a virtual camera is detected
• the camera can’t see a face (or sees more than one)
• sustained talking is picked up by your microphone (the sound itself never leaves your device)

Those are notes for the recruiter to review with context, not automatic penalties. And here is what we never do:

• No video or audio recording. The face check and the talk check run on your own device, and only small notes about what they noticed are sent
• No listening to what you say. We notice that talking happened, never the words
• No photos or biometric data stored, ever
• No reading your screen, files, or anything outside the interview tab

What the AI sees, and what it never sees

The grading model receives your code, the test results, and the question you were asked. Before anything reaches it, we strip personal details like your name, email, school, and graduation year, because none of them have anything to do with whether your code works.

The syntax helper in the interview is there to help you, not to judge you: how much you use it is shown to the recruiter as context, but it never changes your score.

Every score comes with a reason you can read

No black boxes. Every AI score is stored with a plain English explanation of how it was reached, tied to the exact version of the grading rubric that produced it, so it can still be explained months later.

How long we keep things

• 90 days: your code, drafts, and syntax helper chats
• 30 days: proctoring events (tab switches, paste flags, talking flags, and the like)
• 12 months: your score, verdict, and the written reason behind them

After that, it’s deleted automatically, on a schedule.

We check ourselves for bias, every year

New York’s hiring AI law (Local Law 144) requires an annual independent bias audit of tools like this one, and we hold ourselves to it. Same questions, same rubric, same point weights for every candidate on a role. That’s the design, and the audit checks the outcomes.

Your rights, in plain words

• Ask for a different process. You can request an alternative to the AI scored interview.
• Ask why. You can request the reasoning behind your result.
• Talk to a person. A human is reachable at support@recrewit.ai for any of the above.

Your link is yours alone

Your interview link only opens for you: you sign in, and the signed-in account has to match the email your invitation was sent to. Someone who finds your link can’t take your interview or see your work.

Where this applies

The screening service currently operates for candidates in the United States only, and this page describes US practice (including New York City’s Local Law 144).