Skip Navigation

Engineering interview questions for a structured interview, by discipline

Title card reading Engineering interview questions for a structured interview, by discipline, Recruit Engineering

Among the hiring tools compared in a major 2022 review of selection research, the one with the best record for predicting job performance is the structured interview. You have probably also sat through interviews that told nobody anything useful about an engineer. Both are true because the word interview covers two different activities that happen to share a room and a calendar invite.

A chat and a structured interview are different tests

In one version, someone asks whatever comes to mind and the conversation goes well or badly. The interviewer walks out with a feeling. In the other, every candidate gets the same questions about the same parts of the job and each answer is scored against a scale written before anyone walked in.

The US Office of Personnel Management, in its guide to structured interviews, describes the second kind as one that uses rules for asking questions and for observing and rating the answers. It also says that interviews with more structure show higher validity and better agreement between raters, along with less adverse impact.

Validity in this research is a correlation between how people scored when they were hired and how well they later did the job. A score of 0 means the interview told you nothing. In the review by Paul Sackett and colleagues, structured interviews had the highest average validity at a little over 0.4 while cognitive ability tests came in at about 0.3. That makes structured interviews perhaps the strongest predictor of job performance, ahead of cognitive ability, which the field had confidently ranked first for over half a century.

The same authors went back over their results in a 2023 follow-up on designing selection systems. They note that their estimate for unstructured interviews came out roughly 0.2 lower than the older estimate. They also give a wide range for structured interviews, running from under 0.2 to above 0.6.

So the label does not do the work. Two employers can both say they run structured interviews, and one of them can be getting most of the benefit while the other gets only a little. You might reasonably guess that the difference sits in the questions and the scoring, because those are the only parts anyone designs, and they are also the parts you can fix this week.

How to build one for an engineering role

The method is not complicated. It mostly asks you to make your decisions before you meet anybody instead of after, which is harder than it sounds because meeting people is when decisions feel easiest to make.

Step What you do What it looks like for an engineer
Write down what the job needs List the four to six abilities that separate a good hire from a poor one in this specific role For a site civil role that might be grading design, permitting and client coordination
Write questions for each ability One or two questions about past work and one scenario question per ability Ask about a permit that stalled, then pose a utility clash found during construction
Fix the follow-ups Write the probing questions in advance so every candidate hears the same ones Ask which calculations the candidate ran personally
Write the scale first Describe a weak answer and a strong answer for every question before the first interview A strong answer names the evidence behind a design decision
Score alone, then compare Each interviewer scores independently and the panel discusses only the large disagreements The electrical lead and the hiring manager each score before the debrief
Let the scores decide Add up scores by ability and compare candidates on the totals A charming candidate with weak design answers stays weak on paper

The first step is the one that usually gets skipped. OPM calls it job analysis, and the idea is that every question should trace back to something the job actually requires. For engineers the useful version is narrow and concrete: the project types, the equipment, the codes or standards and whether the role needs a license. A list like that rules out a lot of questions you might otherwise ask out of habit.

Past work or hypotheticals

Structured interviews use two kinds of question. Behavioral questions ask what the candidate did in a past situation. Situational questions describe a job-like scenario and ask what they would do. OPM reports that behavioral questions show higher validity for complex, professional and managerial jobs, which is a fair description of most engineering roles.

So lean on past work for experienced engineers and keep a scenario or two as a cross-check. For new graduates, who have a senior design project and maybe an internship to talk about, scenario questions carry more of the load. That is fine as long as the scale says what a good answer contains.

The question that matters most is whose work it was

Engineering is team work, so a candidate can describe a substation or a production line in great detail without having designed much of it. Every behavioral question should have a written follow-up that separates ownership from proximity: which calculations they ran, which drawings they produced and which decisions were theirs. Without that probe you are mostly scoring how well someone describes their former employer’s projects.

Example questions by discipline

These are starting points. Each one should be edited to fit the actual job, and each needs its own scale before you use it.

Civil engineers

For civil engineering roles, ask about the project type and the design responsibility first, then permitting and client coordination. If the job needs a PE license, say so in the posting and confirm it early instead of making it an interview question.

  • Walk through a site or drainage design you were responsible for. Which parts did you design yourself?
  • Tell me about a permit that took much longer than planned. What held it up, and what did you change?
  • Describe a time a client asked for a design change late in a project. How did you handle the cost of it?
  • Scenario: the contractor calls to say your storm pipe hits an unmarked utility line. What do you do that first day?

Mechanical engineers

For mechanical engineering roles, the useful thread is a design taken from analysis through prototype and testing. Name the equipment, materials and loads that matter for your job so you can tell whether the experience transfers.

  • Pick a design you took from analysis to a tested prototype. What did testing show that the analysis missed?
  • Tell me about a part that failed in testing or in the field. How did you work out the cause?
  • Describe a design tradeoff you made between cost and performance. Who disagreed with you?
  • Scenario: your part passes analysis with margin but the supplier cannot machine it at the quoted cost. What do you look at first?

Electrical engineers

For electrical engineering roles, decide first whether you are hiring for power and controls or for electronics, because the questions barely overlap. Then ask about voltage levels, equipment and who owned the design and the commissioning.

  • What voltage levels and equipment have you designed for, and which parts of those designs were yours?
  • Tell me about a system you commissioned or brought up for the first time. What went wrong during startup?
  • Describe a design review where someone found a problem in your work. What was it and what did you change?
  • Scenario: during commissioning a motor trips on roughly one start in three. How do you find out why?

Quality engineers

For quality engineering roles, the main job of the interview is separating inspection work from engineering responsibility. The best material is an investigation the candidate led, the evidence they used to find the cause and the corrective action they verified.

  • Tell me about a nonconformance investigation you led. What evidence pointed you to the root cause?
  • How did you confirm that the corrective action worked, and over what period did you check?
  • Describe a supplier quality problem you resolved with people who did not report to you.
  • Scenario: returns jump for one product line and production says nothing changed. Where do you start?

What a scored answer looks like

A scale only works if two interviewers would give the same answer the same score. Here is one way to anchor the first quality question, on a scale of 1 to 5.

Score What the answer contains
1 Describes the defect and the fix but not how the cause was found, or someone else led the work
3 Led the investigation and names a method, with some evidence for the cause
5 Names the data, explains which causes were ruled out and why, then shows the fix was verified over time

Writing the 5 is the useful exercise, because it forces you to say what good looks like for this job before a candidate who is pleasant to talk to makes the decision for you.

Ways a structured interview drifts back into a chat

  • Interviewers improvise follow-ups, so each candidate ends up answering a slightly different interview.
  • Scores get written after the debrief, when everyone already knows what the hiring manager thinks.
  • A culture fit conversation is added at the end with no questions and no scale, and it ends up carrying the decision.
  • The question bank never changes even after the role does, so the interview keeps measuring the last job.

If you are hiring one engineer this quarter, the next thing to do is write the scale for your four or five most important questions and have a second interviewer read it. If they cannot tell a 3 from a 5, rewrite it before anyone sits down across the table.