Asking Peer Engineers

## Who a peer engineer is on the panel Most loops put at least one engineer in front of you who is neither the hiring manager nor a recruiter: someone at or near your level who would share the codebase, the pager and the review queue with you. They usually do not set the pay band for the posted level, do not own headcount, and often hear the team's roadmap the same week you would. What they do have, and nobody else on the panel has, is a first-hand account of the work you are actually buying into. So the peer round is not a smaller version of the manager round. It is a different question budget, and the way to waste it is to point manager-shaped or executive-shaped questions at someone who can only guess. ## What belongs in this window Six families cover almost everything worth asking a peer: 1. **The path to production.** Ask them to walk one merged change all the way to a user. This is the single densest question in the round: it exposes hand-offs, approvals, ownership and how much of the job is waiting. 2. **Deploy cadence.** How many deploys went out last week, and who pressed the button. Cadence is a proxy for feedback speed and for how big a change has to get before it ships. 3. **On-call load.** How many engineers share the rotation, what the last shift was actually like, and what happens to the work you were mid-way through when you are paged. 4. **Code-review culture.** How long a change waits for a first comment, whether review is a fixed set of owners or whoever is free, and what a disagreement in review looks like when it does not resolve quickly. 5. **Tech debt.** Not whether it exists — it always does — but the mechanism: how a piece of it gets picked up, and whether they can name something that got fixed recently. 6. **Lived experience.** 'What surprised you after you joined?' is the peer-round opener that costs nothing and reaches what the careers page cannot: the tooling nobody warns you about, the meeting nobody mentioned, the part of the job that turned out to be better than advertised. ## What to keep out of it The common miss is aiming org-strategy questions at a peer: next year's bets, how this team maps to company strategy, hiring plans, the reason the role is open, how performance is scored, the pay band. A peer will either recite something they half-remember or say they do not know, and both outcomes burn the window and tell the panel you have not worked out who answers what. Those questions are strong questions — they are just addressed to the wrong person. Save them for whoever on the loop owns them. ## A worked sketch Suppose the role is on a platform team. The peer window opens and you ask three things. First, the walkthrough: what happens to a change after it is approved, from merge to a user. They sketch it — automated checks, a shared pre-production environment, then a promotion step a second team owns. Second, cadence: about 11 changes merged last week but one release cut. Third, the rotation: six people, and the last shift woke them once. (Figures here are illustrative, invented to show the shape of a useful answer.) You have not asked about strategy at all, and yet you now know that the team's feedback loop is a week long, that the author does not control shipping, and that on-call is survivable. That is more decision-relevant than any roadmap answer would have been, and every fact came from someone who could not have faked it. ## The second purpose Your questions are also being read. A peer round usually feeds a written note back to the panel, and 'asked good questions about how we ship' is a line people actually write. Questions that show you have thought about the daily job read as someone preparing to do it; questions cribbed from a generic list read as someone preparing to interview. That is a reason to ask fewer and better ones, and to follow up on the answer instead of moving down your list. ## Practical shape Bring five candidate questions, expect to ask two or three, and pick them live based on what the round already revealed. Ask the walkthrough first if you only get one — it is the highest-yield. Keep a short note after the round: peer answers are the raw material for comparing what different interviewers told you, and disagreement between two of them is more informative than either answer alone.