how it works

One stage, added to the loop you already run.

You configure it once. Your recruiter schedules it like any other interview. A senior engineer runs the room. You get the report the next morning.

01 · onboarding · one call, about a week

Before the first candidate, we spend an hour with the hiring manager and whoever owns the rubric. We leave with four things:

  • Your stack. Languages, frameworks, the shape of the systems the role will touch.
  • Your leveling. What senior means at your company, mapped onto the five dimensions in our report.
  • Two or three signals you most want probed. The thing your last bad hire got past.
  • A domain-shaped exercise. We build it that week. It looks like your work without being your code, and it stays fixed across candidates so the observations stay comparable.

A second hiring team with a different stack or level is a second, shorter calibration. Onboarding is a one-time fee; the numbers are on the pricing page.

02 · scheduling · inside your ATS

The round appears in Greenhouse, Ashby, or Lever as a normal interview stage with its own scorecard. Your recruiter moves the candidate into it the way they would for any internal interviewer. It shows up in your pipeline analytics like everything else.

We confirm a session date with the candidate within 5 business days. Rush and same-day slots exist and cost more. Candidates get a plain email from us that says what to expect and what to bring, and a link to the candidate page on this site.

If you use a different ATS, we schedule by email. It works; it just does not show up in your dashboards.

03 · the session · two to three hours, in person

The candidate arrives at our downtown San Francisco office. The examiner checks a government-issued photo ID against the name you gave us, writes down confirmed or not, and does not keep a copy. Then they sit down together.

The session is a working session, not a quiz. The candidate gets the exercise, a machine or their own laptop, and whatever tools they normally use, including AI assistants. We expect them to use those tools. How they use them is one of the things we are there to see.

The examiner is a senior engineer who has worked in your stack, trained on our fixed assessment spine, and asks the same kinds of questions of every candidate. They take notes throughout. There is no audio or video recording.

What we do not do: trick questions, whiteboard algorithms unrelated to the job, or anything that would make a good engineer feel policed. The in-person format does the verification work on its own.

04 · the report · within 24 hours

You receive a single document, the same one every time. It records the ID check, gives structured observations with a level across five dimensions, includes a short examiner note, and is signed. The code the candidate wrote is attached.

It does not contain a hire or no-hire recommendation. That is deliberate. Your team owns the decision, commercially and legally, and the report is built to make that decision rest on something real.

A 30-minute live debrief with the examiner is included on every retainer, and available on its own otherwise. Read the full sample report to see exactly what arrives.

05 · what stays yours

The decision. The rubric. The candidate relationship. We never contact a candidate about anything other than their session logistics, and we never share one client's observations with another.

Questions we have not answered here are probably on the FAQ. If not, ask us.

The first candidate is free. Book the session.

no contract for the first one · decision-maker joins the debrief