Interviews

How to Prepare for a Technical Interview

A technical interview tests the specialist part of a job by asking you to do some of it. The formats differ between fields, from live coding to a bench test, but the habits that help are shared: understand the problem before you start, explain your reasoning as you go, and deal with being stuck in the open.

Key points

  • Ask in advance which format the technical stage will take, how long it lasts and what tools or references are allowed.
  • Restate the problem and ask clarifying questions before you begin working on it.
  • Explain your reasoning as you work, so the interviewer can follow your method and help if you go off course.
  • When you are stuck, say where, try a simpler version of the problem and accept a hint if one is offered.
  • Be ready to explain every part of a take-home task, a portfolio piece or a project on your resume.

A technical interview checks whether you can do the specialist part of the job by watching you do some of it. You might solve a problem, analyse a case, carry out a practical task or explain work you have already done.

Formats vary a great deal between fields, but what is assessed is usually the same: how you reason and how you work, not only whether you reach the right answer. That is why the habits in this guide carry across fields.

Find out which format to expect

Ask the recruiter or coordinator what the technical stage involves, how long it lasts, what tools you will use and whether you may look things up. It is a normal thing to ask, and employers generally answer. These are the formats you are most likely to meet.

FieldCommon formatsWhat is usually assessed
SoftwareLive coding exercise, system design discussion, take-home task, code reviewProblem solving, clear code, trade-offs, testing habits
Data and analyticsSQL or spreadsheet exercise, case study with a dataset, presentation of findingsAsking the right question of the data, checking results, explaining them to non-specialists
Finance and accountingSpreadsheet modelling test, questions on accounting standards, case studyAccuracy, a model others can follow, stated assumptions
EngineeringWorked calculation, design or drawing review, questions on codes and standardsMethod, safety awareness, sense-checking the answer
DesignPortfolio review, critique of an existing product, short design exerciseThe reasoning behind decisions, response to constraints and feedback
Skilled tradesPractical or bench test, fault-finding task, safety questionsSafe working, method, quality of finish

Clarify the problem before you start

Technical problems in interviews are often underspecified, sometimes on purpose. Starting work on your first reading is a common way to solve the wrong problem. Restate the task in your own words, then ask about anything that would change your approach: the size and shape of the inputs, the constraints, the standard you are working to, and what a finished answer looks like.

Clarifying questions before a data case study

Before I start, can I check three things? Is the aim to explain why renewals fell in the third quarter, or to recommend what to do about it? Does the customer table include accounts that have cancelled, or only current ones? And who are the findings for: the analytics team or the commercial director?

Each of those questions changes the work. A coding exercise has its own versions (How large can the input be? Can it be empty? Should invalid values be rejected or ignored?), and so does a practical test (Which standard applies? Which tools may I use? Is the circuit isolated?).

Think aloud

If you work in silence, the interviewer sees only the final answer and cannot tell a sound method with a slip from a lucky guess. Describe what you are considering, which option you are choosing and why, and what you intend to check. You do not need to narrate every keystroke.

Weaker

Works in silence for four minutes, then says: "OK, I think that works."

Stronger

"I will start with the simple approach: go through the orders once and keep a running total for each customer in a dictionary. That should be fine for the sizes you mentioned. When it runs, I will test an empty list and a customer with a single order."

Why: The strong version tells the interviewer the plan, the reason for it and how it will be checked. If the plan has a flaw, they can say so before ten minutes have been spent on it.

Talking and thinking at the same time is a skill, and it feels unnatural at first. Practise it on two or three problems before the interview, aloud, with a timer running.

What to do when you are stuck

Getting stuck is normal, and some exercises are set to be harder than can be finished in the time. What you do next is part of what is being observed.

  1. Say exactly where you are stuck: "I can find the duplicates, but I cannot see how to do it without sorting first."
  2. Go back to a smaller version: work one example by hand, or write the slow, obvious solution and improve it afterwards.
  3. State an assumption and carry on: "I will assume the beam is simply supported; tell me if that is wrong."
  4. Take a hint when one is offered, and build on it. That is how you would work with colleagues.
  5. If you do not know a fact, say so and explain how you would find it out. Do not bluff.

Coding exercises, system design and take-home tasks

For live coding, practise in the language you will use, and in a plain editor if that is what the interview uses. Get a correct, simple solution working first, then discuss how to improve it. Test it with a normal case and an awkward one before you say you have finished.

System design discussions have no single right answer. Start with what the system must do and roughly how much load it must carry, sketch the main components and how data moves between them, then discuss the trade-offs and what happens when a part fails.

For a take-home task, read the brief twice and respect the stated time limit. Include a short note saying how to run your work, what you assumed and what you would do with more time. If the brief does not say whether libraries, templates or AI coding assistants are allowed, ask. Expect to be asked to explain or extend any part of it afterwards.

Case studies, practical tests and portfolio reviews

Case studies, common in data, finance and consulting-type roles, reward structure. Say how you will break the problem down before touching the numbers, state your assumptions, check that the answer is of a believable size, and finish with a recommendation, not only a figure.

Practical tests in trades, laboratories, workshops and kitchens are assessed on method as much as on the finished piece. Follow safe working practice visibly, ask about unfamiliar equipment before using it, work in a tidy order and say what you are checking as you check it.

Portfolio reviews in design, architecture and writing, and for engineers with project folders, go best with two or three projects you know thoroughly. For each, explain the brief, the constraints, your own part, one decision you would defend and one you would now change. Remove confidential client details, and check what you are permitted to show.

The week before

  • Reread the advert and list the tools, languages, standards or methods it names; revise those first.
  • Go over the fundamentals of your field before anything obscure, because a basic topic can be explored in a lot of depth.
  • Do two or three practice problems against the clock, speaking your reasoning aloud.
  • Reread your own resume. Any project or tool you have listed may be explored in detail, so be ready to say what you did and why.
  • If the interview is remote, test screen sharing and any shared editor in advance; the guide to video interviews covers the setup.

Technical stages rarely stand alone. Expect questions about how you work with other people as well, which the guide to behavioural interviews covers, and see how to prepare for a job interview for research and logistics. Have a few questions of your own ready too, such as how the team reviews work; there are more in questions to ask at the end of an interview.

Common mistakes

  • Starting to solve the problem before understanding it, and finding out ten minutes later that a clarifying question would have changed the approach.
  • Working in silence, which hides sound reasoning and removes any chance for the interviewer to redirect you.
  • Bluffing about a tool, standard or concept you do not know, when saying "I have not used that; here is the nearest thing I have used" is safer and more honest.
  • Chasing a clever solution first and running out of time with nothing that works; get the simple version right, then improve it.
  • Submitting a take-home task you cannot fully explain, or overrunning the stated time limit by a large margin without saying so.
  • Bringing a portfolio of many projects and knowing none of them in depth, when two or three well-explained pieces would show more.

Questions and answers

What if I cannot finish the problem in the time?

Say where you have got to, what remains and how you would complete it. A clear partial solution with a sound plan can still show your method. Check with the interviewer whether they would prefer you to keep going or to talk through the rest.

Is it acceptable to ask for a hint?

Yes. Explain what you have tried first, so the hint can be specific, then use it and carry on. Sitting in silence for several minutes generally tells the interviewer less than asking a precise question.

Can I look things up during a technical interview?

It depends on the employer, so ask at the start. Some allow documentation or a search for syntax, as you would have at work; others want to see what you can do unaided. If it is allowed, say what you are looking up and why.

How should I prepare if I do not know the format?

Ask the recruiter first. If you still cannot find out, prepare the fundamentals of your field, practise explaining two past projects in detail, and rehearse working through an unfamiliar problem aloud. Those three cover most formats.