Data Scientist interview questions
4 questions asked in Data Scientist interviews, from job seekers and our own interviews, each read by us before publishing.
Technical or skills round
How do you choose an evaluation metric for a classification model?
Tip: Tie it to the business cost of errors: precision when false positives are costly, recall when missing a case is costly, and why accuracy misleads on imbalanced data.
How would you explain overfitting to someone who isn't technical?
Tip: Use a plain example, like a student who memorises past papers but cannot answer a new question, then explain how you guard against it: validation sets, cross-validation, simpler models, regularisation.
Hiring manager
Tell me about a project where the data didn't support what stakeholders hoped.
Tip: Show how you presented the finding honestly, with evidence, and suggested what could be done instead.
Assessment or case study
Walk us through a model you built end to end.
Tip: Cover the problem, data, features, baseline, model choice, evaluation, and what happened after deployment. Interviewers listen for the baseline and the business result.