Somewhere in the middle of a hiring loop there is usually a round where nobody senior is in the room. Two or three people who would actually sit next to you get thirty minutes, no rubric, and a vague instruction to see what they think. It sinks more candidates than the technical screen, and almost nobody prepares for it.
The peer round answers a different question
A hiring manager is asking whether you can do the job and whether your level fits the budget. That round has structure, and it rewards the kind of preparation you can plan for. The peer round has neither structure nor a clear rubric, because the people running it are asking something narrower and, honestly, more self-interested: is my week better or worse if this person starts next month.
They will review your work. They will inherit your half-finished handover when you take a holiday. They will sit through your updates every morning for a long time. So they are not measuring peak ability. They are measuring the running cost of working with you, from a thirty minute sample.
The blunt version of their question is: will this person make my job harder. Everything else in that conversation is a proxy for it.
The people asking were never trained to ask
Most peer interviewers found out they were on the panel the day before. Nobody gave them a scorecard. Nobody told them which competencies they own, so they often overlap with the manager and ask you about yourself for the fourth time that day. That has consequences you can use.
They remember feelings, not transcripts. The written feedback that lands in the applicant tracking system is usually three lines typed between meetings, and those three lines are impressionistic: seemed thoughtful, went on a bit, kept talking over me. One vivid moment weighs more here than in a structured, competency-scored format.
Silence means they are deciding what to ask next, not laying a trap. Let it sit for two seconds, then offer something useful rather than filling the gap with nervous restatement.
And the questions will wander. If one is vague, ask what part they are most interested in. That single move reads as collaborative rather than evasive, which is most of the point.
Four things they are reading off you
Whether you actually listen
Do you answer the question asked, or the question you rehearsed on the train? Peers notice interruption far more than managers do, because interruption in an interview predicts interruption in a design review. If you talk over someone once in thirty minutes, they will mention it.
Whether you can say you do not know
The answer they want has two halves. The first is plain: I have not done that. The second is the one candidates skip: here is how I would find out. Name the source, name the person you would go to, name the smallest experiment that would settle it. People who bluff in interviews bluff in incident reviews, and everyone in that room has been burned by it.
How you talk about the team you just left
This is the fastest way to lose the round. Not disagreement, not criticism of a decision, but contempt. The moment you describe your last team as a mess you had to carry, your interviewer stops hearing a story about them and starts hearing a preview of how you will describe this team in eighteen months. Even accurate criticism costs you here, which is unfair and completely real.
Whether they can disagree with you safely
Expect a small, gentle push back on something you said. It is rarely about the technical point. They want to see whether you hold a position with a reason, fold instantly, or go cold. The move that works: ask what they have seen, then say what evidence would change your mind. That is a sentence almost nobody offers, and it lands.
Answer at the altitude of a colleague
Drop the polished candidate register by one notch. Not sloppy, just level. Candidate voice is smooth and strangely general; colleague voice has specifics, timelines and a couple of things that went badly.
Use "we" accurately, and give credit by name. "Our data engineer owned the ingest pipeline, I owned the API on top of it, and we argued about the schema for a week" tells a peer more than any ownership claim, because they do the same job and can smell inflated scope from across the table. The behavioural story structure you use with a manager still applies, but compress the setup and spend your time on the part where something went wrong.
Talk about the boring parts on purpose. How you write a handover document. What you did the Friday a release went out broken. How you review someone else's work when you disagree with the approach but it is not your call. That is the actual texture of the job, and it is what they are trying to picture.
Your half of the hour
This round is also your best and often only chance to find out what the job is really like, because you are talking to the people who live in it rather than the people who sell it. Spend your questions there, not on strategy or compensation, which peers cannot answer and feel awkward being asked.
Three that consistently produce real information:
- What surprised you in your first month here, good or bad?
- What does this team do when it disagrees, and who breaks the tie?
- If you could change one thing about how the team works, what would it be?
Listen to the pause before the answer as closely as the answer. The tie-break question is the most predictive thing you will hear in the entire process: teams that say "we escalate to the manager" work very differently from teams that say "whoever is on call decides and we write it down". If you want more of these, there is a longer list worth stealing from.
The two ways people lose this round
The first is treating it as the soft one. It is often scheduled last, at four in the afternoon, after a day that has already drained you, and the temptation is to coast. Half-answers, an anecdote you have told three times that day, eyes drifting to the clock. Peers read that as disinterest in them specifically. If you have a full day booked, protect your energy for the back half of it.
The second is the opposite: treating peers as rivals. Correcting them on their own stack, explaining their product back to them, dropping seniority, or addressing your answers to whoever looks most senior in the room. That last one happens constantly in rounds with several interviewers at once, and the people you ignored are the ones writing feedback.
Here is the structural fact that makes both mistakes expensive. Peer feedback is usually a veto, not a vote. A peer almost never gets you hired on enthusiasm alone; the offer comes from the manager and the budget. But any single one of them can sink you with one sentence, and "I would rather not work with them" is a sentence hiring managers take seriously, because overriding it costs them a relationship they need every day.
What to do the week before
Ask your recruiter who is in the round and what they do. Read whatever those people have published, not to flatter them, but so your examples land in their world rather than yours.
Then prepare four things. One story where you were wrong and someone junior corrected you. One where you changed your mind because of a colleague, with the colleague named. One sentence about your last team that you would be happy for them to overhear. And three questions for them, written down, because at 4pm you will not invent them.
Last piece of housekeeping: peers usually skim your CV ten minutes before the call, straight from the calendar invite, so it has to make sense to someone who is not a recruiter. If yours is still a raw export of your profile, a tool like Postulit will turn the LinkedIn version into a document that reads like a CV. Fix the first half of the first page and stop there.
If you walk out of a peer round able to describe what that team does when it disagrees, you got something out of the half hour no matter what they decide about you. That is a useful way to hold it, and it also happens to make you much better company in the room.