Electrical Engineer Interview Questions 2026

Last updated: October 5, 2026

By Tom Kenaley, President and Senior Partner, KORE1

The electrical engineer interview questions that work in 2026 make a candidate produce a number, mark up a schematic, and explain when they would refuse to work on something energized. Definitions of Ohm’s law and power factor tell you who studied. They don’t tell you who can do the job.

An automotive supplier outside Greenville, South Carolina, hired a senior hardware engineer two summers ago off a loop that was, honestly, a pretty good textbook exam. Thévenin equivalents. The difference between a BJT and a MOSFET. Explain impedance. He aced every bit of it. Then the first prototype board for a seat-heater controller landed on his bench and he plugged it straight into a 12-volt supply with no current limit. Two regulators cooked. Nobody had asked him what he does in the first ten minutes with a new board, because nobody on the panel had thought of it as an interview question. Four months later the role was open again.

I don’t blame the panel. They asked the questions every list on the internet tells you to ask.

Those lists are mostly written for candidates, which is fine, except the answers are printed right under the questions and every serious applicant has read them. So a question like “what is the difference between AC and DC?” doesn’t test anything anymore. It tests reading.

What follows is the question set we hand clients when an electrical engineering staffing search reaches the panel stage. It sits next to two pieces we’ve already published, so a few things are deliberately missing. Our guide on how to hire an electrical engineer covers the five sub-disciplines, the loop structure, and the project walk-through question. The electrical engineer salary guide owns the comp bands. This page is the questions. Nothing else.

Why Textbook Questions Stopped Working

Two reasons, and the second one is newer.

The first is the one above. Popular question lists print model answers, candidates memorize them, and the interview turns into a recitation contest that rewards whoever prepared hardest the night before. That’s a real skill. It’s just not electrical engineering. You end up hiring a strong student and then discovering, around month three, whether you also hired an engineer.

The second is that remote first rounds now come with a second screen. A candidate who can’t explain a buck converter can have a decent paragraph about one in front of them in about four seconds, and it’ll be correct. Definitions are free now. What isn’t free is a specific number for a specific circuit, produced on the spot, with the reasoning out loud, or a red pen moving across a schematic nobody has seen before. Those still have to come out of a person.

The Bureau of Labor Statistics put the median electrical engineer’s pay at $120,630 in May 2025 and projects 8% growth in electrical and electronics engineering jobs from 2025 to 2035, with about 16,300 openings a year, according to the Occupational Outlook Handbook. That’s a lot of hiring. A miss costs real money at that salary, and it costs a lot more in schedule. For a local read on what an offer should look like, the KORE1 salary benchmark tool is quicker than any survey.

Run the Napkin Round First

We borrowed this years ago from a hiring manager at an engineering firm in Irvine, back when the hardest seat we had open was a switchgear designer for a hospital campus. She’d slide five questions across the table on the back of a lunch receipt. Every one had an answer a non-engineer could check, which sounds like a gimmick until you remember that the person running round one is often a business-side manager or an HR partner who couldn’t tell a good derivation from a confident one if their bonus depended on it.

No calculator. Good ones round hard and tell you they’re rounding. The others freeze. Or recite a formula for a while and never arrive anywhere.

You won’t use all five on one person. Two from their world, one from next door. Your utility candidate ought to crush the motor row and the wire row, and if they look at the 75-ohm load and says “honestly, no idea, I’d start with the reflection coefficient and go from there,” great. I’ll take that over a bluff every day of the week. The grade is on getting from words to a number. And on admitting where your map ends.

One thing to watch for in the third row. A lot of candidates forget the return conductor and give you about 6 volts. Then the good ones catch it on their own about thirty seconds later. Catching it counts for more than never missing it.

Power Systems and Facility Electrical Questions

These are for utility, data center, industrial facility, and building electrical roles, the ones closest to the bottom of the frequency range. If the seat is in that world, our power systems engineering recruiters run these on the first call, so by the time a candidate reaches your panel they’ve heard a version of them once already. Good. They should still answer cleanly a second time.

“A fault on a branch circuit tripped the main instead of the feeder. What happened, and what do you look at first?”

You want the word coordination inside the first two sentences. The upstream device was faster or more sensitive in that current range than the one closest to the fault, so the whole building went dark for a problem on one circuit. A strong engineer goes to the time-current curves next, then asks whether a breaker setting got changed during maintenance, which, in our experience, is the honest answer more often than a design error. Bonus if they mention that selective coordination is a code requirement for some systems, emergency and legally required standby among them, and not just a nice-to-have.

“The utility just told you the available fault current at your service went up. What do you redo?”

Short answer from a good candidate. “The short-circuit study, then the arc flash study, then the labels.” Medium answer adds that equipment interrupting ratings need checking, because a breaker rated for 22 kA is now maybe sitting on 30. The best answer adds that the arc flash assessment has to be reviewed at least every five years anyway, and a utility change is exactly the kind of event that should trigger it early.

Follow up with IEEE 1584. The 2018 revision of the arc flash calculation guide added five electrode configurations, VCB, VCBB, HCB, VOA, and HOA, and changed how enclosure size feeds into incident energy, according to the IEEE Standards Association. Anyone who has run a study in SKM or ETAP since then has had to pick an electrode configuration for every bus. Ask which one they used for a typical 480 V MCC and why. VCB is the usual call. An engineer who did the work remembers choosing.

“What does an arc flash label tell you, and what doesn’t it?”

Underrated question. It tells you incident energy at a working distance, the arc flash boundary, and the PPE category or rating. It doesn’t tell you the study was right, or that it’s current, or that the breaker upstream still has the settings the study assumed. People who’ve been in the field long enough have found a label from 2014 on gear that got a new transformer in 2021. They’ll tell you about it. Usually with some feeling.

Power Electronics and Analog Questions

Middle of the frequency range. Switch-mode supplies, gate drives, motor control, op-amp front ends, the places where thermal behavior decides whether the design survives a real duty cycle.

“Your buck converter works fine on the bench and fails at 85 °C ambient. Walk me through it.”

A Houston client building drives for downhole tools used this one on four finalists last spring, and it sorted them instantly. Two said “check the thermal design” and stopped. The third talked about MOSFET on-resistance rising with temperature, which is true, roughly 1.5 to 2 times higher at hot. The fourth went straight to the inductor. Ferrite saturation flux density drops as it heats, so a part with comfortable margin at 25 °C can saturate at 85, the current spikes, the FET takes the hit, and things go downhill quickly. Then she asked what the inductor’s saturation rating was specified at, and whether anyone had checked it at temperature. They hired her the following week.

Was the third guy wrong? Not really. His FET point holds up fine. But she had a ranked list of suspects in her head and he had one, and on a hot board at 2 in the morning that’s the whole difference.

“Why might a converter pass every functional test and still fail radiated emissions?”

Somebody who has actually failed a scan will talk about the switch node, that little copper island flipping between ground and the input rail a few hundred thousand times a second, and about the input loop and the cables hanging off the board doing a decent impression of an antenna. Everybody in this line of work has a chamber story. It always ends in a re-spin. I like asking how many re-spins, and then what the last one cost, because people round the cost down and the count up, and you learn a lot from which way they round. The ones who’ve been burned now pre-scan on a near-field probe before booking a chamber, since the second booking is never on a date anyone likes.

“When would you pick a linear regulator over a switcher, on purpose?”

Flip side of the regulator row. Some people learned “switchers good, linears bad” in school and never revisited it. Ask anyway. You’re hoping to hear about a low-current rail, a little dropout, a sensor front end or a PLL that couldn’t stand any ripple, and ideally the actual part number they put there, because people remember the part number when it was their call.

PCB and Hardware Design Questions

“A new board lands on your desk for first power. What do you do in the first ten minutes?”

Greenville never asked it. Should have.

What you want, more or less in this order. They look at it under a loupe first, paying special attention to diodes, electrolytics, and anything someone hand-reworked at 6 p.m. the night before. They ohm out each rail to ground cold. They power it from a bench supply with the current limit turned way down, bring rails up one at a time if the board lets them, and watch the current and touch things, or point a thermal camera at it, long before firmware is even in the conversation. Some will mention a fuse. Some will mention smelling it. Both fine.

“Plug it in and flash it.” Seven years of experience, some of them. The good ones confess to the time they skipped a step and what it cost, usually without you prompting them, and usually with a wince.

“Why put a 100 nF and a 10 µF capacitor on the same rail, and when does that combination hurt you?”

Part one is easy. Bigger cap handles the lower frequencies and the small one handles the higher ones, so together they keep the rail quiet over a wider band. Most people get there.

Part two separates them. Put two caps with different self-resonant frequencies in parallel and you can get an anti-resonant bump in between, right where some clock harmonic happens to live, so the extra capacitor you added “just to be safe” is the reason the rail rings. An engineer who has stared at an impedance plot knows. One who copied the eval board doesn’t.

“Tell me about a board you re-spun. What would you have caught with one more day of review?”

All of them have one. Fine. Did it leave a scar in the form of a specific habit? One candidate in Tempe told a client he now runs design rule checks against the board house’s real capabilities, not the CAD defaults, after a 3-mil trace came back open on every unit of a 200-piece build. That’s an answer. “Better communication with layout” isn’t one.

Firmware-heavy seat? Different loop. Interrupts, memory, and RTOS live in our embedded software engineer interview questions, and PLCs, loop tuning, and machine safety live in the automation engineer interview questions. Left out of this page deliberately.

RF Questions for the Top of the Spectrum

Honest admission. I’m the wrong person to grade an RF answer, and probably so is most of your panel, which is why we tell clients to borrow an RF engineer from a sister team or a consultant for one hour rather than nod along to answers nobody in the room can check. Three things we’ve watched those borrowed engineers ask:

• Sensitivity came in 3 dB worse than the link budget. Where do you look? Front end first. The first stage’s noise figure sets most of the chain, and anything lossy in front of the LNA counts against you directly, a filter, a switch, the six inches of coax someone added in week nine.
• How do you know a VNA cal is good? You measure something known right after. An engineer who has never had a bad cal hasn’t done much measuring.
• What was different between the ADS or HFSS model and the first board back? Everybody has one of these. The good ones name the assumption that broke.

Deeper RF seats go through our RF engineer recruiters, who run their own screen.

Hand Them a Schematic With Three Mistakes in It

If a client gave me twenty-five minutes and said pick one exercise, it’d be this. We call it the redline. One printed page and a red pen.

Take a real schematic from your own archive, a simple one, a page with a microcontroller, a regulator, a relay output, and a sensor on I²C. Plant three or four errors. Ours usually include:

• A relay coil switched by a transistor with no flyback diode across it.
• An I²C bus with no pull-up resistors anywhere on the page.
• A non-rail-to-rail op-amp on a single 5 V supply, expected to swing all the way to ground.
• A reset pin left floating.

Hand it over. “This came back from a design review. What would you mark?” Then leave the room for ten minutes if you can. People think better without someone watching.

Three things get scored. How many of the planted errors they found, obviously. Which one they found first, because the flyback diode is the one that destroys hardware and engineers with field time tend to see it in under a minute. And what they asked before marking anything. The best candidates almost always ask a question about operating conditions first. What’s the relay’s duty cycle? Is this running off a battery? How long is the cable to that sensor? Those questions aren’t stalling. That’s what reviewing a design actually looks like.

For power systems roles, swap the schematic for a one-line diagram. Same idea. A feeder breaker sized larger than the main above it, a transformer secondary with no overcurrent protection, a generator transfer switch with no indication of how the neutral is handled. A facility engineer who has stamped drawings will find at least two of those quickly, and will probably also find a fourth thing you didn’t plant. That happens more than you’d expect, and it’s a very good sign.

Safety Questions Where Seniority Shows

NFPA 70E, the standard for electrical safety in the workplace, has a 2027 edition that NFPA released in 2026. One of its changes is small on paper and big on a job site. Per Electrical Contractor magazine’s review of Article 130, when an energized electrical work permit is required and specifies shock or arc flash PPE, at least one additional person trained in emergency response now has to be present, outside the limited approach or arc flash boundary, whichever is greater. The permit itself also moved, from 130.2(B) to 130.3.

You don’t need a candidate to recite article numbers. Nobody does that in real life. You do want someone who works near energized equipment to know the edition changed and roughly what changed.

“When is it acceptable to work on something energized?”

Rarely, and with paperwork. De-energizing would create a greater hazard, or it’s genuinely infeasible, or it’s diagnostic work like voltage testing that can’t be done any other way. Then a permit, a risk assessment, the right PPE, and now the second person. A candidate who answers “if you’re careful” is telling you something important about how they’ll behave at 2 a.m. on a plant shutdown.

We placed a facility electrical engineer at a semiconductor fab expansion in Chandler, Arizona, earlier this year, and the client’s panel had one other finalist. Better résumé, honestly. Eleven years, a strong napkin round, a clean one-line redline. He said he’d work a live 480 V bucket “with the right gloves” to avoid shutting down a tool line. The client’s EHS manager was in the room. That was the end of that.

“Who is allowed to open this panel at your current site, and how do you know?”

Qualified person, in 70E terms, means trained and knowledgeable for the specific equipment and task, not just a job title. Good candidates describe a real system, a training record, a list, a lockout program under OSHA 1910.147. Weak ones say “the electricians.”

“Tell me about the last time you stopped a job.”

Everyone who’s spent time near energized work has stopped one. If they haven’t, either they’re very junior or they don’t say no easily. Neither is disqualifying. You just want to know which.

Judgment Questions for Senior Engineers

For a senior or staff seat. At that level a bad hire rarely costs you a wrong calculation, somebody catches those, so it costs you a wrong call that nobody caught until the field returns started coming in.

“What’s something you signed off on that you’d hold today?”

The good ones barely pause. They’ve been carrying it around. A capacitor run at 80% of rated voltage because the schedule was slipping. A single-source PMIC that went end of life eighteen months after launch and forced a board redesign nobody budgeted for. A vibration test a vendor talked them out of. I always want the follow-up too, which is what they check now that they skipped back then, and the best answers to that are boringly specific.

“The simulation says one thing and the bench says another. Which do you believe?”

Bit of a trick. “Neither, yet” is right. Plenty of bench numbers are wrong because of a 6-inch ground lead on the probe or a bandwidth limit somebody left switched on, and plenty of models are wrong because nobody put the parasitics in. You want a person who keeps poking at both until one explains the other.

“When did a standard tell you no, and what did you do about it?”

UL, IEC, the NEC, some customer’s 140-page spec. Everybody runs into one. Some engineers redesign to fit, some find a compliant workaround, and a few go argue that the clause doesn’t apply to them. That last group is occasionally right. When they’re wrong it tends to cost a quarter.

Weak Answers and What to Ask Next

None of these rule anybody out alone. Each is a reason to push once more before scoring.

One weak answer is an off moment. Three in an hour is a pattern.

What Good Candidates Ask You

The questions coming back across the table are part of the interview too. Strong electrical engineers tend to ask some version of these:

• What’s on the bench, and is it shared? An engineer who asks about scope bandwidth or a spectrum analyzer has spent real time in a lab.
• Who owns compliance testing, and how many chamber days are budgeted?
• Who stamps the drawings here? (Facility and utility roles.)
• How many re-spins did the last product take?
• Who do I go to when the schematic and the layout disagree?

If a candidate asks none of these and only asks about remote days, it doesn’t mean they’re wrong for the job. It does mean you haven’t seen how they think about the work yet.

What EE Hiring Managers Want to Know Before the Panel

How many technical questions fit in one interview?

Eight to ten technical questions fit in a one-hour electrical engineering interview, as long as each one gets a real follow-up.

More than that and you’re running a quiz. The good stuff is in the second question, the “why that part” and “what was plan B” and “what did the scope actually show.” I’ve sat in on panels that got through twenty-two questions in sixty minutes and walked out knowing about as much as a written test would have told them, minus the paper trail.

Should an electrical engineer get a take-home test?

Rarely. A 25-minute schematic redline in the room tells you more than a weekend take-home, and senior engineers won’t do the take-home anyway.

We used to like them. Then every take-home started coming back looking oddly similar, tidy and correct and a little generic, and a client in Austin finally asked one finalist to walk through his submission live. He couldn’t explain the snubber he’d drawn. Nobody can outsource marking up a page they’ve never seen while talking a stranger through it. If something absolutely has to go home, send a sanitized schematic of yours to review, never a blank-sheet design problem.

Do these questions work for entry-level engineers?

Yes, with the napkin round and the redline as the core and the judgment questions mostly dropped.

A new grad hasn’t stopped a job or signed anything off, so skip those. They can absolutely get to 3.5 watts, though. A surprising number of them spot the flyback diode too, usually the grads whose capstone was a real thing with a real board rather than a simulation and a poster. I always want to hear about that capstone. What broke on it, and who fixed it at midnight.

If a candidate has a PE license, can we skip the technical round?

Don’t skip it. A PE license proves someone passed a demanding exam and has supervised experience, not that they know your equipment or your codes.

The NCEES PE Electrical and Computer exam is split into Power; Electronics, Controls, and Communications; and Computer Engineering, and the Power exam alone runs 80 questions over eight hours. That’s real. It’s also a different test from whether this person will coordinate your switchgear correctly. Keep the napkin round short for a PE, then spend the time on the one-line redline and safety.

How do we stop candidates from using AI during a remote interview?

You mostly can’t on video, so move the parts that matter, the napkin round and the redline, into the room or onto a shared page where you watch the pen move.

Video’s fine for the get-to-know-you part. Motivation, why they’re looking, what they want next. For the technical bits we push clients to bring people in, and when that’s impossible, to use a shared whiteboard and have the candidate narrate as they go. Narrating is hard to fake. A pasted answer shows up whole, with no false starts and no crossed-out arithmetic, and once you’ve seen that a couple of times you start spotting it in about five seconds.

Who on our team should run the redline?

The engineer who’ll review the new hire’s first design should run it, because they already know which mistakes matter in your products.

Hand them a one-page rubric. Three columns. Errors found, which came first, what got asked before the pen moved. And if there isn’t a senior EE in the building, which honestly is often the reason the req exists, that’s the round to borrow one for, whether that’s a consultant, a former employee, or a recruiter who screens technically.

Write Down the Answer You Want Before You Ask

Last suggestion, and it costs nothing. Before the panel meets the first candidate, every interviewer writes down the answer they’d accept for each question. Two lines. A number, a reason.

It sounds bureaucratic. It’s actually the fastest fix for the most common panel failure we see, which is four engineers walking out of a debrief with four different ideas of what a good answer was, then picking whoever they liked.

Our recruiters average more than 15 years in the seat, KORE1 has been placing engineers since 2005, and 92% of our placements are still in the job a year later. A big share of that comes from questions like these being settled before the first candidate walks in. Whether you need a direct hire or a contract engineer for a project that ends, our electrical engineer recruiters can bring a screened shortlist and the redline page. If a search has stalled, talk to an engineering recruiter and we’ll look at it with you.