Last updated: November 16, 2026
By Jennifer Burdick, Recruiting Manager, KORE1
The best director of engineering interview questions in 2026 ask the candidate to name the managers they led and defend an org chart and a delivery date, because the job happens through those managers. Questions about architecture and leadership style can stay in the loop as a warm-up. They are not enough.
Director of engineering reqs reach my desk at KORE1 in two kinds, the first search and the repeat one. The repeat search costs more, and it usually starts the same way. A hire who interviewed beautifully has been in the seat for seven or eight months, and the managers underneath are leaving. The managers saw it first.
Take a composite from the last couple of years, with names and some figures changed. A logistics software company in Providence, Rhode Island, about ninety engineers in five teams, hired a director at $230,000 base after a four-round loop. The candidate had led seventy engineers at a larger company. He drew a clean system diagram on the whiteboard, described his leadership philosophy for ten minutes without a stumble, and had a ready answer for every question about vision and roadmap. It was a good diagram. Eight months later two of the five engineering managers had resigned, a third was taking recruiter calls, and the platform migration had slipped twice. The debrief was short. Nobody on the panel had asked him to name a manager he had promoted, a manager he had removed, or what he did the week one of them stopped hitting dates. Not one question reached the layer where the job lives. Replacing the two managers took close to five months, and the migration restarted under a new owner. Expensive. Avoidable.
That gap is what the rest of this page works on. A candidate can rehearse a leadership philosophy for an hour. Nobody can rehearse a name, a date, and what happened next, not twice, and the follow-up questions on a real story are what separate a person who has managed managers from a strong engineering manager who has watched someone else do it. The difference shows fast. Most published lists run thirty or forty questions about vision, culture, and communication. Nearly every one has a prepared answer waiting. Candidates know it.
I will name my conflict before you find it. KORE1 earns nothing on a director search until the person we put forward signs and starts, and a director who leaves at month seven means we run the search again. A loop that works matters to us as much as it matters to you. Our director of engineering staffing desk runs versions of these questions in screening calls, ahead of any client panel.

A Philosophy Can Be Rehearsed. A Name Cannot.
A director of engineering is accountable for the delivery, structure, and health of three to six engineering teams, and does that work through the managers who lead them. The unit of work moves from the sprint to the quarter. The unit of feedback moves from the code review to the manager who is not growing. Different job.
Nothing here replaces the rounds you already run. Keep them. The rung below is covered in our engineering manager interview questions, and the rung above, with its CEO relationship and company-wide arithmetic, is in the CTO interview questions. If the seat leans toward infrastructure and vendors, the IT director interview questions fit better. This page spends its hour on the middle layer, which is the one most loops skip. The whole search these questions sit inside is laid out in our guide to hiring a director of engineering.
Demand for people who can do it is not the problem. The Bureau of Labor Statistics files directors under computer and information systems managers, so there is no separate director count, but that category projects 16 percent growth from 2025 to 2035, roughly 53,500 openings annually, and median pay of $175,140 as of May 2025. Directors who have really managed managers get recruited constantly. A slow loop loses them.
Five jobs fill the week. Each one leaves evidence behind, and this table pairs the job with the evidence and the question that reaches it.
| The job | Evidence that exists only if they did it | Where the question starts |
|---|---|---|
| Growing managers | First names, conditions, outcomes | Name the managers who reported to you |
| Shaping the org | What moved, and what broke when it did | A printed org chart |
| Delivering across teams | A date with odds attached | Four teams, one shared deadline |
| Spending headcount | The annualized cost, not this year’s | Fourteen requests, one budget |
| Translating upward | The last bad news, and how many days it sat | The day a committed date moved |
Make the Candidate Produce the Managers
Start here. Stay longer than feels polite. Pick two of the three. Not all of them.
Name the managers who reported to you at your last job, and start with the one who surprised you.
A director who has done this answers in specifics inside thirty seconds. A first name, the team, the shape the manager was in when the candidate inherited them, and what happened over the next year. Who got promoted. Who moved sideways. Who left, and whether the candidate saw it coming. Then pick one name and keep going. What did you tell her in the first month? What did she get wrong that you let her get wrong?
The weak answer arrives as a philosophy. “I believe in empowering my managers” is a sentence with no person in it.
Fifteen minutes on this one question is worth more than an hour on style. I mean that literally. A 2015 Gallup analysis found that managers account for at least 70 percent of the variance in employee engagement scores, which is why a director who cannot develop managers is a risk to every team beneath them, even in a quarter when the dates hold.
One of your managers hits every date. Half of her team has asked to transfer. What do you do on Monday?
Strong candidates refuse the premise that a team hitting its dates is a healthy team. They talk to the engineers before they talk to the manager, and they want to know what the dates are costing, in weekend commits, skipped reviews, and the one person quietly doing the work of three. They look at the transfer requests for a pattern. All from one project, or from one meeting? Then they say what they would tell her, which is usually harder than it sounds, because she has done what the company rewarded. That is the hard part.
A weak answer congratulates her and offers coaching on communication. Next question.
Your strongest manager wants your job. She is not ready. What do you say, and when?
Listen for timing and for specifics. The best answers say it early, in plain words, with two or three things she would have to show first, and treat the not-yet as information she is owed instead of as a kindness to be managed. I have watched finalists answer this one with a promotion date. Nobody should promise a date. The answer that costs a candidate the offer is “I would never let her feel that,” because she will feel it either way, and the only choice is whether she hears it from the director or from a rumor.
An Org Chart With Five Problems in It
Print this one. Hand it over and say the VP wants a single change before the quarter ends. Names are made up.
| Team | Manager | Engineers | Also true |
|---|---|---|---|
| Payments | Priya | 14 | Owns the billing service |
| Platform | Lars | 3 | Also the tech lead and the only person on call for the cluster |
| Mobile | Vacant | 7 | A senior engineer has been “acting” for nine months |
| Growth | Dan | 8 | Shares the notifications service with Payments |
| Data | Marc | 6 | Reports to the director, but Finance sets its roadmap |
You can move one person and redraw one line. What do you change, and what breaks?
There are five. Most candidates find three. Fourteen engineers under one manager is too many for real one-on-ones, since our own director of engineering page puts a manager’s span at five to ten. Platform has a manager doing an engineer’s job and carrying the pager alone. One pager, one person. Mobile has had an acting manager for nine months, with the authority and the career path of neither role. Growth and Payments both own the notifications service, and a service with two owners has none. And Data answers to the director while taking direction from Finance.
Melvin Conway’s old observation, on his own site, that any organization designing a system will produce a design whose structure copies the organization’s communication structure, is the reason the shared service matters, and a good candidate sees the problem without ever naming the law.
Order matters more than count. Another composite, from a payments company in Chandler, Arizona, that ran a chart like this in a final round last spring. The finalist who got the offer spent her first four minutes asking what the VP had been asked to fix, circled the vacant Mobile seat before the span of fourteen, and said she would move nobody until she had talked to the person who had been acting for nine months. The other finalist redrew the whole chart into six boxes. Confident. Tidy. A reorg costs a quarter of everyone’s attention, and the second answer spent it without asking what for.

Four Teams Report Green and the Date Is Still a Coin Flip
The VP wants the new billing platform by the last day of the quarter. Four teams own pieces of it. In the weekly review each team says it is on track, with about 85 percent confidence.
What is the chance the date holds, and what do you tell the VP?
Multiply. Eighty-five percent four times over is about 52 percent, which is roughly even odds, and that assumes the four teams are independent of each other. They are not. Mobile cannot start until the Payments API ships, so a slip in one team moves the next. A strong candidate does the multiplication out loud or says “about a coin flip” inside ten seconds, then says what they would do with it. Cut scope. Pull one dependency forward. Give the VP a range this week instead of a surprise in week eleven.
The weak answer says four green teams make a green project. It trusts the color of the status and never asks what the color is made of. Status colors are opinions.
Your VP loves this slide. What does it say?
| Quarter | Deployments per week | Change lead time | Change fail rate | Failed deployment recovery time |
|---|---|---|---|---|
| Q1 | 38 | 3.1 days | 9% | 4 hours |
| Q2 | 47 | 2.4 days | 13% | 7 hours |
| Q3 | 61 | 1.8 days | 19% | 11 hours |
The numbers are the exercise’s. Real dashboards are messier. What matters is the shape. Two columns got better and two got worse, so the group ships faster and more often while failing nearly twice as often and taking almost three times as long to recover. The DORA metrics guide splits software delivery performance into throughput and instability for exactly this reason, and treats the two sides as a pair.
A strong candidate asks what changed in the second quarter. Was there a push to ship, or a cut to testing, or a rule that everything reaches production within two days? They ask what the failed deployments cost customers. And they say the slide should not reach the board with only the flattering half on it. Weak answers celebrate the velocity.

The Headcount Plan Nobody Annualized
One page, handed over cold. Your managers have asked for 14 new engineers. The budget for the year is $2.4 million. A loaded engineer costs $225,000 in this exercise, an assumption for the page and not a benchmark.
Fourteen requests, $2.4 million, one year. What is your answer?
The tempting arithmetic divides $2.4 million by $225,000 and funds ten. That is the trap. The better answer starts with a question about start dates. Hires who land in July cost half a year, so fourteen of them run about $1.58 million this year and fit comfortably. The same fourteen cost $3.15 million the following year, and that second number is the one nobody put on the slide.
Strong candidates say “it fits this year and breaks next year” without a prompt. They go on to propose a phasing, a priority order tied to the roadmap, and which managers get told no, and they are ready to say it to those managers directly. Weak candidates fund ten at random, or ask for more budget. Both are wrong. Our look at engineering org chart ratios is useful background for panelists who have never built one of these plans.
Bad News Moves Upward on a Schedule
Ask when the last piece of bad news went up the chain. Then ask how many days it sat in the candidate’s hands first. Ask what they said to the VP on the day a date they had committed to moved, and then ask what they said to their managers the same afternoon. Listen for a day count. Directors who have lived through a few reorgs measure delay in days and hold themselves to the number. Directors who have not tend to say “as soon as I had the full picture.” That sentence means three weeks. Days, not feelings.
One more worth keeping in your pocket. A product leader says, in front of your managers, that engineering is the bottleneck. What do you say in that room, and what do you say afterward in private? The first answer shows whether the managers can trust the director to stand in front of them. The second shows whether the product leader can.
Call the Managers Who Used to Report to Them
The reference that matters is not the old VP. It is a manager who reported to the candidate, and our director searches make that call on every finalist. A few questions that work on it:
- What did you know about your own job in month six that you did not know on day one, and who told you?
- When did they last overrule you, and were they right?
- Tell me about the worst week you had under them. Did they know?
- Would you take another job under them tomorrow? Two seconds of silence tells me more than the answer.
- Of the managers who reported to them, how many are still doing the job?
That is the whole call. Ask the candidate for three former managers who will take that call. A candidate who cannot produce three has already answered the first question on this page.
Fixer or Scaler: Where the Hour Goes
Five exhibits do not fit in an hour. They should not. Choose by the situation you are hiring into.
| The situation | Spend most of the hour on | Keep short |
|---|---|---|
| Missed dates and managers leaving | The managers, then the delivery forecast | The headcount plan |
| A healthy group that is about to double | The headcount plan, then the org chart | Bad news upward |
| A reorg or an acquisition just landed | The org chart, then the managers | The headcount plan |
| A new VP who wants to see everything | Bad news upward, then the managers | The org chart |
Put three people in the room. The VP who will own the hire. One engineering manager who would report to the candidate, because that person will recognize a real answer to the manager questions from the first sentence and will ask the follow-up you did not think of. And someone from finance for the headcount page. Keep it to four rounds over about two weeks. Most of our director searches run 40 to 70 days from kickoff to signed offer, and a loop that drags past three weeks is a common place to lose a finalist. Fast loops win.
If You Are the One Being Interviewed
Some people reading this are on the other side of the table, so here is the short version. Bring names. Before the first round, write down every manager who reported to you, what shape they were in when you started, and where each one is now, and be ready to say what you got wrong with at least one of them. Bring one org change with numbers on it, including what it cost. Numbers, not adjectives. Practice the arithmetic in the delivery and headcount exercises until you can do it out loud without a calculator. And ask your interviewers the same questions. The director who asks “who are the managers, and how long has the Mobile seat been vacant” in round one is doing the job already.
Worries We Hear From the VP Before the First Interview
Should a director of engineering still be tested on system design?
Briefly, and as a reader instead of an author. A director should be able to look at an architecture, find the risk, and say who needs to be in the room, in about twenty minutes. A forty-five minute whiteboard design is testing the job they left, and it rewards the candidate who stayed closest to it.
How many rounds should the loop run?
Four rounds, spread over about two weeks. Run a conversation with the VP, a panel that includes one engineering manager, the exhibit round with finance in the room, and a final with whoever the director will lean on most. Four is enough. A fifth round almost never moves the decision. It does change whether the finalist is still answering your recruiter’s emails.
Can we promote our best manager instead of running a full search?
You can, and the same loop applies. Hand your internal candidate the org chart, the forecast, and the headcount page like anyone else. The exercise is not an insult. It is the first honest conversation about what the seat requires, and a manager who handles it well has just earned the promotion in front of everyone, while one who stumbles has learned something before it cost either of you a year.
What does a director of engineering cost in 2026?
Base pay for a director typically runs $190,000 to $280,000 at US employers, and total compensation climbs well past that at funded startups and large technology companies once equity lands. Stage, industry, and metro move the number hard. Equity is the swing. Our director of engineering salary guide breaks bands out by tier, and the salary benchmark assistant returns a current range for your metro.
Director or VP, how do we know which seat we are interviewing for?
Count what reports to the seat. If the direct reports are engineering managers and the seat answers to a VP, it is a director, and if the seat owns the whole function and answers to the CTO or CEO, it is a VP. Interviewing for the wrong one is expensive in both directions. Know which before you schedule. The questions on this page assume the first, and a VP of engineering search needs a different set.
What if the strongest finalist has never managed managers?
Then you are hiring a manager and paying for a director. Sometimes that is the right call, because some excellent senior engineering managers are one rung from the seat and ready to grow into it, but price it that way, say it to them plainly, and plan to coach. I would also ask whether the better move is to fill the seat below first through engineering manager staffing and let the group tell you who is ready. Tangent, and a real one, but the search for the wrong rung is the commonest way these go sideways.
Hold the Offer Until a Former Direct Report Calls Back
Keep only one thing from this page, and it should be the call. Before the final round, get one engineering manager who reported to the finalist on the phone and ask the five questions above. Whatever you cannot learn in ten minutes is your next interview.
Once someone signs, the first quarter decides most of what follows, and we wrote about that in an engineering leader’s first 90 days. If you want help building the exhibits from your own org, bring the org chart and the delivery plan to our engineering leadership recruiters and we will build the exercises and the shortlist from them. Director seats run as direct hire, and you meet three or four finalists, not a stack of resumes. Retention is the number I would ask any agency for first. Ours is 92 percent at twelve months across the people we place.

