Engineering manager interview topics and how to prep for them

An engineering manager behavioral interview is an evidence hunt. The interviewer usually has a short list of leadership competencies to score and about an hour to find proof for or against each one. They are not grading the story. They are trying to find the moment where you, specifically, made a call that a weaker manager would have avoided or delayed.

Every follow-up is them digging for that moment. That is why experienced managers come out of these rounds saying "I knew the answer, I just couldn't say it." The decision was in there, buried under four minutes of context about the reorg, and the interviewer ran out of time before they found it.

So prepare decisions, not stories. A story is what happened. A decision is the point where you chose one path over another and knew what it would cost. Interviewers can score a decision. They struggle to score a narrative, and they can't score "we" at all: your team shipped the migration, but they want to know what would have gone differently without you.

Most EM behavioral questions, whatever the wording, are a version of one of the six topics below. For each one you'll see how it tends to be asked, what the interviewer is listening for and where answers usually go wrong. The general behavioral interview questions page covers the individual-contributor versions.

A list of fifty questions trains you to recognize questions. It does not give you anything to say. Work from your own material instead.

1. Write down eight to ten situations from the last few years as one-line notes. Situations, not answers: "the payments API dependency in Q3", "Priya's performance plan", "the reorg when we lost two seniors".
2. Map each situation to the six topics above. Aim to cover every topic at least twice. Gaps are where you will blank, so find them now rather than in the room.
3. For each situation, write the decision sentence: "I decided to ___ instead of ___ because ___." If you can't fill it in, the story is not ready, however good it feels.
4. Write down the cost. Who was unhappy, what slipped, what you gave up. Answers without a cost sound rehearsed, because real management decisions always have one.
5. Find the result you can say out loud: a number if you have one, an observable change if you don't. "The two teams now share an on-call rotation" is a result.
6. Prepare for the second and third question, not just the first. "What would you do differently?" "What did your manager think?" "How did you know it was working?"
7. Say each one out loud, timed, before the interview. Aim for about two minutes on the first pass. Rambling is invisible on paper and obvious out loud.

Keep each story to one or two topics in a single loop. Panels often compare notes in the debrief, and hearing the same migration from three interviewers makes your range look narrow. The STAR answer worksheet walks through turning one situation into a structured answer step by step.

The six topics still apply at Anthropic, OpenAI and other frontier labs. What shifts is the material the interviewer expects in your stories. Tension between research and engineering priorities is a common theme, and so is a trade-off between shipping speed and safety or quality. Have at least one story about each, and be ready to talk about why you want to work on the lab's mission specifically, in your own words.

Anthropic publishes guidance on how candidates may use AI during its hiring process. It encourages using Claude to research the company and practice answers, and asks that live interviews be done without AI assistance unless the interviewer says otherwise.¹ Practicing out loud on your own is the preparation that holds up in that room.

For round-by-round detail, read the Anthropic engineering manager interview guide or the OpenAI engineering manager interview guide.

Don't open another question list. Write your eight to ten situations as one-line notes and map them to the six topics. Pick the topic with the thinnest coverage and write one decision sentence for it. Then say that answer out loud once, timed, and listen for the point where you started explaining context instead of what you did. That point is where tomorrow's practice starts.