Top 50 model training Interview Questions

Explain practical ways to train and evaluate a classifier when the target classes are highly imbalanced.

model trainingSupervised LearningClass ImbalanceIndeedVerisk AnalyticsAtos

Implement a CART decision tree from scratch, including split selection, stopping rules, prediction, and evaluation.

Feature Engineeringmodel trainingSupervised LearningAgile DefenseState StreetPublicis Groupe

Handle rare positive labels in ad fraud detection with the right sampling, loss design, validation, and thresholding strategy.

Feature Engineeringmodel trainingClass ImbalanceNextrollAmexAmerican Family Insurance

Describe a machine learning project, justify model and evaluation choices, and explain how you addressed its main engineering challenges.

Hyperparameter TuningCross-ValidationFeature EngineeringBookingPAPERKLA

Handle severe class imbalance in rare failure prediction while balancing recall, precision, and operational alert volume.

Feature Engineeringmodel trainingClass ImbalanceRivianBaker HughesAvathon

Tests ownership during an ML production failure, including diagnosis, cross-functional communication, and learning from offline-vs-production gaps.

production failuremodel trainingDebuggingIntuitive SurgicalUSTGGigapower