Interview Questions to Ask a Data Scientist Candidate
Questions that surface real evidence of models and analysis that actually drive a business decision, not just interesting findings — plus what to actually listen for in the answer, not just whether they say the right buzzwords.
1. You are asked to predict customer churn, but only a third of the records have a reliable outcome label. How do you proceed?
What to listen for: Expect checking label quality, sampling bias, simple baselines first, honest uncertainty, and a plan to collect better data before trusting complex models.
2. Your model scores very well in testing but performs poorly after launch. What do you investigate?
What to listen for: Look for leakage, drift between training and live data, evaluation on realistic splits, and monitoring that would catch the gap early.
3. A manager wants a model that uses a sensitive personal attribute because it improves accuracy. What is your view?
What to listen for: The candidate considers fairness, harm, and policy, proposes alternatives or audits, and explains the trade-off without lecturing the manager.
4. You have three analysis requests and time for one. How do you choose, and what do you tell the other two requesters?
What to listen for: A strong answer ties choices to decisions the business will actually make, estimates effort, and communicates delays with a clear alternative or timeline.
5. What would you do in your first 30 days in this data scientist role?
What to listen for: Whether their plan is grounded in the role's actual must-haves (statistical modeling and python/r) or generic ("get to know the team"). Specificity here signals real understanding of the job, not just interview prep.
6. Why are you leaving your current role, and why this one specifically?
What to listen for: Forward-looking motivation tied to something concrete about this role, versus badmouthing a past employer or vague dissatisfaction. Watch for whether their reasons actually match what this role offers.
A warning sign in this interview
Reaches for the most complex model by default, cannot explain results to non-technical people, or treats a high accuracy number as proof the model works.
What to look for on the CV before the interview
Models shipped, the measured impact and the data size worked with.
Confirm before you make an offer
A quantitative degree is common. Ask for work that reached production.
People who may apply under another title
This job is also called Machine Learning Engineer, Applied Scientist, Data Analyst (Advanced), Research Scientist. A candidate with one of those titles can be a strong match, so ask the same questions and judge the answers, not the title.
Frequently asked questions
How many of these questions should I ask in one interview?
3-4 is realistic for a 30-45 minute interview once you leave room for follow-ups and the candidate's own questions. Divide competencies across panel members rather than every interviewer asking all of them — see the Interview Scorecard tool for a structured version of this.
How is this different from the Interview Scorecard tool?
This page is a generic starting point for the data scientist role. The tool generates a full panel-ready kit — competency-grouped questions, a 1-5 rating rubric, and strong/weak answer examples — tailored to your actual job description and competencies.
What is a red flag when interviewing a data scientist?
Reaches for the most complex model by default, cannot explain results to non-technical people, or treats a high accuracy number as proof the model works.
What should I confirm before offering a data scientist the job?
A quantitative degree is common. Ask for work that reached production.
Hiring a data scientist? The rest of the toolkit
- How to write the job description →
Titles, must-haves, credentials and a starting posting.
- Search strings to find candidates →
Five searches from broad to narrow, with exclusions.
- Run the whole hire in one place →
Jobs, CV review, interviews, offers and checks together.