Frequently asked questions

Practising senior and staff engineers who have been through our review process: an application, a technical review, and a calibration session with a member of our team. They are approved for named interview types rather than in general — being a strong backend engineer does not automatically qualify somebody to run a system design interview.

Offers go out within the hour, and an interviewer typically accepts the same day. The scheduling then depends on your candidate: we show you the interviewer’s real availability, so the first bookable slot is genuinely bookable.

We tell you why, in terms you can act on: nobody holds every required skill, or nobody is approved for that interview type, or people match but none are free in the window you gave us. The remedies are different, so the message is different.

Yes. Interviews are scored against a rubric attached to the role, and you can see it before anyone is interviewed against it. What you cannot do is change the rubric for an interview that has already started — the evaluation has to stay reproducible.

Only where you want us to. Most teams keep the hiring manager conversation and the team-fit stages in house, and hand over the parts that are structurally identical every time — the coding exercise, the system design discussion.

Every interview is evaluated on its own evidence. We do not show an interviewer a previous evaluation of the same candidate before they submit their own, for the same reason we do not show them a colleague’s.

An interview plus its write-up. You set your own availability and weekly capacity, and you are offered work that fits inside both — never outside them.

Always, and without giving a reason. If you do give one it feeds our matching, so that a run of “too little notice” tells us a company’s scheduling policy is wrong rather than that you are unreliable.

The fee settles when the evaluation is submitted, not when the interview ends. The work is the interview and the write-up, and paying for half of it before the other half exists would be the wrong incentive.

They see your first name, your surname initial, your headline and your experience. Not your full name, not your rate, not your contact details, and never the private notes you take while the interview is running.

No. No model rates a candidate, writes an evaluation, summarises what was said, transcribes the call, or infers anything about the person. Every word in an evaluation was written by the interviewer who ran the session, and it is auditable back to them.

No. The meeting link is a placeholder for whatever conferencing tool you already use, and we neither join nor record it.

A Next.js application over PostgreSQL. Interview evidence is stored against the rubric it was collected under, so an evaluation written in January still shows January’s questions even if the rubric has been edited since.