Whenever you're ready, walk me through how you'd approach this.
Why did you choose one model over another, and how do you approach feature selection and hyperparameter tuning including regularization like Lasso and Ridge?
Answer this as a practical modeling exercise. Explain how you would compare candidate models, prevent information leakage, select or remove features, tune regularization strength, and choose an operating point using validation data. Include a reproducible scikit-learn implementation and discuss how your approach changes for classification versus regression.