← All posts
general

Amazon Interview Debrief: How the Hiring Decision Is Made

You finish the last interview of your Amazon loop, close the video call, and wait. Somewhere in the next few days, the four to six people who each spent an hour with you meet for the Amazon interview debrief, and that meeting decides the outcome. You are not in the room. By the time it starts, every interviewer has already written down what they think of you and voted.

Most advice treats the loop as the interview and the debrief as paperwork. It is closer to the other way round. The loop collects evidence; the debrief judges it, and the judges work from written notes, not from how the conversation felt. Once you know how that reading happens, you answer every question in the loop differently. Amazon does not publish its debrief procedure. What follows is the picture that interviewers who have sat in debriefs and candidates who have been through them consistently describe, with the places it varies marked as such.

What the debrief is

The debrief is the meeting where the interviewers from your loop compare their written feedback and reach a hire or no-hire decision, together with a view on level. Three things about it are consistent across accounts, with details that vary by team.

  • Who attends. The interviewers from the loop, the hiring manager (who is usually one of them), and the Bar Raiser. The recruiter often joins to take notes and handle logistics, but does not vote. Phone screen feedback is sometimes read into the record; the phone screener rarely attends.
  • When it happens. Usually within a few days of the last interview, sometimes the same day for an in-person loop. The gap stretches when interviewers travel, when the loop was split across two days, or when someone's feedback is late. Candidates report waiting anywhere from a couple of days to a few weeks for a decision, and most of that is scheduling and approvals rather than deliberation.
  • The format. Written feedback first, then discussion. By most accounts, each interviewer has to submit their write-up and vote before they can read anyone else's, and the Bar Raiser will not start the meeting until everyone has. The meeting itself is often under an hour. A loop where the feedback agrees can be quick; a split loop takes longer.

The debrief decides more than yes or no. It decides the level the evidence supports, and occasionally whether to run one more interview because a principle was not covered or two interviewers heard contradictory stories.

The written feedback

Before the loop, each interviewer is assigned two or three of the leadership principles to cover. Their write-up addresses those principles with evidence from your answers: what happened, what you specifically did, what resulted, and what you would do differently. Many interviewers type near-verbatim notes during the interview, so the write-up quotes you, with the interviewer's assessment kept separate from the record of what you said.

Each write-up ends with a vote. Accounts describe it as a four-point inclination: strongly inclined to hire, inclined, not inclined, strongly not inclined. The exact wording differs between accounts, and some describe a hire or no-hire call with a strength attached, but the shape is the same: a direction plus a confidence. The vote comes with a level judgment, because "inclined at the level below" is a common and legitimate outcome.

When Amazon interviewers say "data", they mean something that can be checked for consistency: a number with a baseline, a date, a named system, a decision with the alternatives you rejected. "Improved performance significantly" is not data. "Cut p99 latency from about 800 to 300 milliseconds by moving the join into the cache layer, measured over two weeks" is. Interviewers are asked to write evidence rather than impressions, and they cannot write down what you did not say.

That is why vague stories die here. An interviewer who liked you but has only "seems strong on ownership" in their notes will be asked in the debrief what the evidence was. If they cannot point to a specific, their vote softens or the Bar Raiser treats the principle as uncovered. The room gives the benefit of the doubt to the evidence, not to the candidate.

How the discussion runs

The Bar Raiser typically runs the meeting; on some teams the hiring manager runs it with the Bar Raiser acting as a check. Either way, the Bar Raiser's job is to keep the discussion on written evidence rather than on who liked the candidate.

By consistent accounts, the discussion goes roughly like this:

  1. Everyone reads everyone else's feedback, silently, at the start or shortly before. Votes are visible to the group.
  2. The group walks through the principles or the interviewers in turn. Where two interviewers heard the same story, they compare the details: who owned the decision, what the number was, what the candidate said they did personally.
  3. Interviewers challenge each other's evidence. "You wrote that they owned the migration; they told me the migration was led by their tech lead." A contradiction between rounds is the most common way an inclined loop turns into a no hire.
  4. The Bar Raiser asks the level question: would this person raise the bar for the level, meaning be stronger than the current average Amazonian at that level in that role, not merely good enough to fill the opening.
  5. Votes can change. People update after hearing evidence they did not have. The decision is reached by consensus, and by most accounts both the hiring manager and the Bar Raiser need to agree for an offer to go ahead. The Bar Raiser can stop a hire the manager wants; they cannot force one.

The point for you is that the written record travels. A specific you give one interviewer becomes evidence for all of them, and a contradiction you create between two rounds is found in minutes.

Down-levelling and split decisions

A down-level offer comes about when the evidence supports a hire but the scope of your stories matches a lower level than the one you applied for. A candidate interviewing for a senior role whose strongest stories are about owning a component and influencing their own team is, in the debrief's eyes, a hire at the level below. Whether that turns into an offer depends on the team having an opening at that level and the hiring manager wanting to make it; the recruiter will usually ask whether you would consider it. Candidates report this most often in senior and principal loops, where the level bar is less visible from outside.

Split votes, say two inclined and two not, are resolved by discussion first. The Bar Raiser pushes the group to name what the disagreement is actually about, which is usually one principle or one contradicted story. From there, accounts describe three outcomes: the group converges once the evidence is laid side by side; an extra interview is arranged to cover the gap, which candidates sometimes experience as a surprise follow-up call; or the loop declines. The culture, as interviewers describe it, is to decline when in doubt. A no hire with clearly documented reasoning is treated as a good outcome for the process, even though it is a bad day for the candidate.

What makes feedback read well

Interviewers write the feedback, but you supply the material. The write-ups that carry a debrief share a few properties, and all of them are in your control.

  • Consistency across rounds. If you tell the same story to two interviewers, the role you played, the numbers and the result need to match. Interviewers compare notes, and your resume and phone screen are part of the record too.
  • Numbers with baselines. "Reduced cost by 30 percent" invites "from what, and how did you measure it?" Give the before, the after and the method without being asked.
  • Your own actions, in the first person. "I decided", "I wrote", "I pushed back". The interviewer has to separate what you did from what happened around you, and "we" makes them guess.
  • Honest failure stories with a result. A failure with a real cause, a real cost and a change you can show in a later story reads as strength on Earn Trust and on growth. A failure that is secretly a success reads as evasion.
  • Depth that survives follow-ups. The follow-up chain is where evidence is generated. If you cannot go a level deeper on your own story, the interviewer writes "could not explain how the number was measured", and that line is read aloud. The Dive Deep guide covers what that depth looks like.
  • Scope that matches the level. Stories about tasks and weeks for a senior role, or about organizations and years for an entry role, both create a level discussion you do not want.
  • Short openers. A ninety-second answer leaves time for the questions that produce evidence. A six-minute monologue leaves the interviewer with a summary and no quotes.

What you can do before the debrief

After your last interview, almost nothing. Feedback is often written within hours, sometimes before you have left the building or closed the call. By the time a thank-you note arrives, the votes are in.

  • The thank-you note. Send one if you want to; it is polite and it is harmless. Send it to the recruiter, since you usually cannot reach interviewers directly. Do not use it to add evidence or correct an answer. By most accounts it does not enter the written record and nobody reads it in the debrief.
  • Writing to the recruiter. Fine for a timeline question. The useful conversation with the recruiter happens before the loop: which level you are being considered for, which principles the loop tends to cover, and whether there is a written exercise. Ask those early.
  • The real lever is preparation. Build a story bank of eight to ten stories, each mapped to two or three principles, and check them against the questions by principle. For each story, write the follow-up chain yourself: baseline, measurement, who else was involved, what you rejected, what you would change. Tell each story the same way twice. Then practice with someone who interrupts, because the written feedback is produced under follow-up pressure and solo rehearsal does not create any.

The STAR method gives the answer shape; the debrief is why the Result part needs a number and the Action part needs the first person.

The wait and the outcome

Candidates report hearing back anywhere from a few days to two weeks after the loop, with longer waits for senior levels, for loops that triggered an extra interview, and when approvals sit with someone on leave. Silence is usually scheduling, not a signal. A polite check-in with the recruiter after a week is normal.

The outcome arrives through the recruiter. A positive decision comes as a call with a verbal offer, then the written one, and sometimes with the level question attached if the loop decided on a lower level than you applied for. A negative decision usually comes as a short note or call. Detailed interview feedback is rarely shared; Amazon's general practice, as candidates describe it, is not to pass on the interviewers' notes, though a recruiter will sometimes name the broad area that fell short.

On reapplying, recruiters commonly tell declined candidates to wait before interviewing again for a similar role, and the periods candidates report are in the range of six months to a year. Amazon does not publish a fixed rule, it varies by role and by how close the loop was, and a decline at one level does not bar you from a different role or team. If you are declined, the single most useful thing to do is write down every question you were asked and every follow-up you struggled with while you still remember them. That list is the preparation plan for the next loop, wherever it is.

Where coaching fits

The closest you can get to seeing your own debrief is a mock interview with someone who has sat in them, who writes feedback in the same form, evidence per principle with an inclination and a level, and then reads it back to you line by line. Six of the coaches on techinterview.coach have Amazon on their history and run exactly that kind of session; you can see them at /coaches/company/amazon. If you want a zero-cost first pass, the free practice interview runs a behavioral round with an AI interviewer and gives you a written recap of what carried evidence and what did not. It is free; sign in with LinkedIn.

Related guides