25 Data Scientist Interview Questions

Walk me through a full data science project — from problem definition to stakeholder presentation.

How do you decide whether a business problem needs machine learning or a simpler analytical approach?

Describe a time your analysis led to a surprising or counterintuitive conclusion.

How do you explain p-values and statistical significance to a non-technical executive?

What's your approach to exploratory data analysis on a new dataset?

How do you handle missing data — what's your decision process?

Describe your experience with causal inference vs. correlation analysis.

How do you design an experiment when you can't randomize assignment?

What's your approach to feature selection for a predictive model?

Describe a time your model worked well in development but failed in production.

How do you communicate uncertainty in your findings without losing stakeholder confidence?

What's your experience with time-series forecasting?

How do you validate that a machine learning model is actually improving a business metric?

Describe your approach to building a recommendation system.

How do you handle datasets that are too large to fit in memory?

What's your experience with natural language processing?

How do you ensure your analysis is reproducible and auditable?

Describe a stakeholder who was skeptical of your data-driven recommendation. How did you handle it?

What's your approach to detecting and handling outliers?

How do you prioritize which analyses to do when there are too many questions?

Describe your experience with clustering and unsupervised learning.

How do you build a data science function from scratch in a company that hasn't had one?

What's your experience with Bayesian methods?

How do you approach bias detection in models used for people decisions?

What's the difference between a good data scientist and a great one?