Best interview questions to ask candidates

The 10 best interview questions to ask candidates

Each is explained further down. Ask four to six, leave time for the candidate's own questions, and keep the set identical across candidates for the same role.

1. Walk me through a project you owned end to end — what was your specific contribution?
2. Tell me about a technical decision you made that turned out to be wrong.
3. What part of your last job did you find genuinely difficult?
4. How would you approach [a real problem from this role] in your first month?
5. Describe a time you disagreed with someone more senior. What happened?
6. What is something you changed your mind about in the last year?
7. How do you decide what to work on when everything is urgent?
8. Tell me about something you shipped that you are not proud of.
9. What would your last manager say you needed to work on?
10. What do you want to be doing in this role that you are not doing now?

Questions about work they have actually done

These are the highest-signal questions you can ask, because the candidate cannot prepare a generic answer and the details are checkable. The goal is to get past the team's accomplishments and find the person's own.

• “Walk me through a project you owned end to end.” Follow with *what did you personally build?* Strong answers name specific decisions, constraints and trade-offs. Weak answers stay in the first person plural — “we decided,” “we shipped” — and never narrow to one person's work.
• “What was the hardest part, and why that part?” Good candidates pick something genuinely hard and explain the difficulty precisely. A candidate who says the hardest part was a framework they had not used before is telling you the project was not hard.
• “What would you do differently if you started it again?” This separates people who reflect from people who narrate. An answer with no regrets at all is usually a sign of shallow ownership, not flawless execution.

Questions about failure and judgement

Everyone has a rehearsed failure story. The useful move is to ask for a decision rather than an incident, because decisions reveal how someone reasons under uncertainty.

• “Tell me about a technical decision you made that turned out to be wrong.” You are listening for what information they had at the time, when they noticed, and what they did next. “I would have needed data we did not have” is a fine answer; “it was the right call, the requirements changed” usually is not.
• “Tell me about something you shipped that you are not proud of.” Candidates who can name real technical debt they created, and explain why they accepted it, are describing engineering judgement rather than perfectionism.
• “What is something you changed your mind about in the last year?” An inability to answer is itself signal. People who learn from their work can usually name a belief that did not survive contact with it.

Questions about how they work with other people

• “Describe a time you disagreed with someone more senior.” Look for whether the disagreement was resolved with evidence, escalated appropriately, or simply dropped. Candidates who have never disagreed with anyone senior are either very junior or not telling you the truth.
• “How do you decide what to work on when everything is urgent?” Strong answers describe a mechanism — asking what breaks if this slips, checking with whoever owns the deadline. Weak answers describe a feeling.
• “What would your last manager say you needed to work on?” This is a better question than “what is your greatest weakness” because it is attributed. It is harder to answer with a disguised strength.

A forward-looking question that is worth the time

Give the candidate a real, slightly simplified problem from the role and ask how they would approach it in their first month. You are not looking for the right answer — you do not have one either. You are watching how they scope an unfamiliar problem, what they ask about before proposing anything, and whether they can say “I would need to find out X first” without discomfort. This is the closest a conversation gets to watching someone do the job.

Questions to drop

• Brainteasers and puzzles. They measure exposure to puzzles. Google published its own findings on this years ago and stopped.
• “Where do you see yourself in five years?” It rewards candidates who are good at telling you what you want to hear, and penalises honest ones.
• “Why do you want to work here?” as a scored question. Worth asking for the conversation, but scoring it mostly measures who had time to read your careers page.
• Anything about protected characteristics — age, family plans, nationality, health, religion. These are unlawful in most jurisdictions regardless of intent; see interview questions you cannot legally ask.

Ask the same questions to everyone

A structured interview — the same core questions, in the same order, scored against the same rubric — is one of the few things in hiring with consistent evidence behind it. It is also what makes your interviews comparable: if every interviewer asks whatever occurs to them, your debrief is a conversation about impressions rather than evidence. Write the questions down before the loop starts and put them in a scorecard.

Where Interviewer Mode fits

Disclosure: we make CoPilot Interview, which includes Interviewer Mode for the hiring side and a real-time assistant used by candidates. We have an interest in you using the first, and our view of the second is that interviews should be designed to measure something an assistant cannot fake. Weigh the advice below accordingly.

Frequently asked questions

What are the best interview questions to ask candidates?

Open questions about work the candidate has actually done: walk me through a project you owned end to end and what you specifically contributed, a technical decision that turned out to be wrong, what part of the last job was genuinely difficult, and how they would approach a real problem from this role in their first month. Ask the same set to every candidate for the role.

What interview questions actually predict job performance?

Structured, job-related questions scored against a rubric predict performance considerably better than unstructured conversation. Work-sample questions about what the person has really done, and realistic problems from the role, carry the most signal. Brainteasers and culture-fit chat carry very little.

How many questions should I ask in an interview?

Four to six substantive questions in a 45- to 60-minute round, with time reserved for the candidate's questions. Fewer questions explored deeply produce more signal than ten asked shallowly, because follow-ups are where the real information is.

Should I ask every candidate the same questions?

Yes for the core set. That is what makes candidates comparable and is the main mechanism that reduces bias in interviewing. Follow-up questions should differ, because they depend on what the candidate actually said — the structure applies to the questions you plan, not to curiosity.

What questions should I avoid asking candidates?

Anything touching protected characteristics such as age, family plans, nationality, religion or health, which are unlawful in most jurisdictions. Also drop brainteasers, hypothetical personality questions, and anything you would not score consistently across candidates.