Interview Practice Questions

What model do you usually use as baseline?

How is boosting trained?

How to select and tune key model parameters?

Can model overfit and how to detect it?

What clustering methods do you know?

What classification metrics do you know?

In what cases can you use recall?

How to recover lost data or information in ML project?

How to compare ML model with simple rules or heuristics?

How to measure churn rate reduction with ML model?

How to determine customer churn propensity?

Difference between theoretical ML knowledge and practical application?

Why is one metric insufficient to evaluate ML model effect?

How to organize logging and monitoring process for ML models?

How is ML used for cross-selling and additional service recommendations?

What can happen to logistic regression when assumptions are violated?

What does out-of-time validation mean in ML models?

How to identify two main global causes in ML problem?

How to tune hyperparameters to squeeze maximum from model?

How to use internal metrics for ML model validation?