Whenever you're ready, walk me through how you'd approach this.
How do you diagnose and mitigate overfitting in high-dimensional models?
Discuss a practical workflow for detecting overfitting, separating it from data leakage and distribution shift, and selecting mitigation techniques. Address validation design, regularization, feature selection, model complexity, and the metrics or plots you would use to verify that the solution generalizes.