Explain Train-Test Performance Gap
Company: ByteDance
Role: Data Scientist
Category: Machine Learning
Difficulty: easy
Interview Round: Technical Screen
Overview: This question evaluates a candidate's competency in diagnosing model generalization failures for both classical machine learning and deep learning systems, including the ability to differentiate true overfitting from issues such as covariate shift, concept drift, data leakage, poor dataset splitting, label noise, class imbalance, and metric or threshold mismatch. It is commonly asked in the Machine Learning domain to assess both conceptual understanding and practical application of model evaluation and validation techniques, probing how a practitioner reasons about robustness, evaluation trade-offs, and appropriate mitigation approaches without focusing on implementation details.