Data Science interview questions and answers

Selection bias means the sample does not represent the population you care about.

You only analyze winners still visible — failed cases disappear.

Clear metric, sample size, runtime, and guardrails.

Volume, null-rate, and distribution monitors with alerts.

Lead with the decision, then evidence, then caveats.