Modelling techniques

Supervised models interview questions

This is the largest category, because supervised learning is still most of the job. Interviewers rarely want a textbook definition of linear regression or a random forest. They want to know when you’d reach for each one, what it assumes, and what you’d expect to go wrong.

Expect questions comparing linear and tree-based models on the same data, asking what multicollinearity does to coefficients, why bagging and boosting behave differently, and how you’d explain a logistic regression’s output to someone who isn’t a data scientist. The better answers tie each model’s behaviour back to the shape of the data.

Try one now

Friendly

Linear regression and a decision tree are both fitted to the same data. When would you expect each to win, and why?

Free, no sign-up. The interviewer asks follow-ups on your reasoning, then rates your answer.

What the 36 questions cover

  • Linear regression17
  • Bagging vs boosting7
  • Linear vs tree-based models, when to choose7
  • Naive Bayes5
  • Random forests5
  • Gradient boosting (XGBoost, LightGBM)5
  • Multicollinearity3
  • Generalised linear models2
  • Logistic regression2
  • Decision trees2
  • Quantile regression1
  • k-nearest neighbours1
  • Support vector machines1
  • Pruning1
  • Stacking and blending1
  • Multi-class strategies (one-vs-rest, one-vs-one)1
  • Multi-label classification1

A question can cover more than one topic, so the counts add up to more than 36.

How the practice works

  1. The AI interviewer asks the question and follows up on your answer, the way an interviewer would.
  2. Stuck? Ask for a hint. Each one gives a little more away, so you can get unstuck without being handed the answer.
  3. Get a written debrief. Four ratings, each from Needs work to Interview-ready, tied to what you actually said, and what to study next.

More modelling techniques questions

See all categories