Data and features interview questions
A model is only as good as the columns it’s given, so interviewers spend real time here. Expect to be handed a prediction problem and asked which features you’d build, then which of them you couldn’t actually have at the moment the prediction is made.
Data leakage runs through this whole category: a feature that quietly encodes the answer, a scaler fitted on the full dataset, a target encoding computed before the split. Interviewers also ask how you’d handle missing values, outliers and high-cardinality categories, and what each choice assumes about why the data looks the way it does.
Try one now
A delivery app wants to predict how long an order will take to arrive. What features would you use, and which couldn't you use at prediction time?
Free, no sign-up. The interviewer asks follow-ups on your reasoning, then rates your answer.
What the 17 questions cover
- Data leakage11
- Train/serve skew5
- Missing values3
- Feature encoding3
- Outlier detection and treatment2
- Feature selection2
- Feature scaling (standardisation vs min-max)1
- WOE binning and information value1
- Feature crossing1
A question can cover more than one topic, so the counts add up to more than 17.
How the practice works
- The AI interviewer asks the question and follows up on your answer, the way an interviewer would.
- Stuck? Ask for a hint. Each one gives a little more away, so you can get unstuck without being handed the answer.
- Get a written debrief. Four ratings, each from Needs work to Interview-ready, tied to what you actually said, and what to study next.