Live Interview · Assessment
D. Okafor — Senior IT Infrastructure Engineer
Summary
Claim Risk Score
40 / 100
Run and extend an existing hybrid identity estateYes
Design an identity migration from scratchNo
AxesDepth 3 · Problem 2 · Ownership 2 · Communication 4
What the score means
0–24The claim in the CV does not holdRISK
25–49It holds, but for a narrower roleNARROW
50–74Holds at the level of the role, with gapsFIT, with caveats
75–100The claim is confirmedFIT
The score rates one claim from the CV, not the person. It weighs ownership heaviest and communication lightest: overstating “led the migration” costs a candidate far more here than plain English does. Confidence says how much one hour could establish at all — at L we are not allowed to write FIT or a score above 60.
Why NARROW
Real operational experience, but not at the level the job description assumes. Hire him to run and extend an estate somebody else designed and he will do it well. Hire him to plan your identity migration and you are betting on something he has not actually done.
What we probed
The CV claims: “Led the migration of 1,400 users from on-prem Active Directory to Entra ID, zero downtime.” That is the largest claim on the page, so most of the hour went into it. No puzzles, no trivia — one claim, pushed until it either held or stopped holding.
Three follow-ups did the work: what broke during the cutover, why Entra ID rather than staying hybrid, and who made the call.
Every quote below carries a timecode into the call transcript. A report without quotes is not a report — we reject it and do not pay for it.
Ratings
Technical depth
3 / 4
Strong where he actually lived. Described a hybrid join failure in detail — the symptom, how they narrowed it down, what it did to the helpdesk queue, and the fix. That is not something you can read up on.
07:12 — “About sixty machines never finished hybrid join and for two days we had no idea why. The OU they sat in had never been added to the sync scope. The fix was one checkbox. Finding it cost me two days.”
Problem solving
2 / 4
Weakest axis. Asked why full cloud identity rather than staying hybrid at their size, he gave the vendor answer — modern authentication and lower maintenance — and could not describe the trade-off in his own terms or name what they gave up. When pushed on what he would do differently, he restated the same plan.
11:40 — “Honestly the design came from the consultancy before I was involved. I inherited the plan.” — volunteered, not extracted. Credit for that.
Communication
4 / 4
Explains cleanly, distinguishes what he knows from what he assumes, and says “I don't know” without padding. Would be reliable on an incident bridge and safe in front of an end user who is already angry.
Ownership
2 / 4
The gap between the CV and the conversation sits here. “Led the migration” turned out to mean he owned the endpoint side of an already-decided migration and ran it well. The identity design, the conditional access policy and the cutover plan belonged to an external consultancy that has since rolled off.
Flags
- “Led” is doing heavy lifting on the CV. He did not oversell it in conversation — he corrected it himself within two questions — but the written claim would mislead anyone who did not ask.
- “Zero downtime” is not what happened. Asked directly, he said finance lost access for about forty minutes during the cutover. He volunteered it without hesitation, so this reads as CV shorthand rather than deception — but it is written on the CV as an absolute.
- No design experience under constraint. Everything he described was executed inside a decision someone else made.
- Single-employer depth. Four years, one estate, one scale. Not disqualifying, but he has not had to be wrong about an environment yet.
If you proceed
- Give him a design exercise with a real constraint — “three hundred users, half on shared shop-floor machines that cannot do MFA, no budget for new hardware.” His reasoning under constraint is the open question, and one hour could not close it.
- Ask him to name a change he argued against and lost, and what he thinks now. His answer to ownership questions will either hold or repeat.
- Take the reference from the consultancy lead, not from his manager. The manager saw tickets closed; the consultant saw judgement.
- If the role is “run and extend”, most of the above is irrelevant and he is a good hire. Re-read the job description before you decide.
What this assessment cannot tell you
- An hour tests two or three claims deeply, not the whole CV. The rest of it is untested.
- We do not assess culture fit, motivation, or whether he will stay. We were not in the room for that.
- We did not watch him work in a live console. This is about depth of experience, not tool fluency.
- We take a transcript of the call, never a recording, and delete it 30 days after delivery. The quotes above stay yours.
- The candidate receives this same document, unedited, on the day it reaches you. An assessment its subject is not allowed to read is one we should not have written.
- A NARROW verdict is not a rejection. It means the evidence supports a tighter role than the one advertised.
- We do not score IQ, culture fit, or whether someone will stay. None of that is measurable here, so none of it is in the number.
- We publish no market medians until we have twenty assessments for the same kind of role. Percentiles on a small sample are false precision.
Assessed by
KAGE-01
Amazon Bar Raiser and IT support manager
One of the trained interviewers Amazon gives veto power over a hire, and a manager who ran an IT support team and hired for it. On a paid assessment the client receives this expert's real name, photo and LinkedIn before the call is scheduled — the codename protects people who are currently employed, not their identity from you.