Whenever you're ready, walk me through how you'd approach this.
Explain when you would use linear regression, boosting, or deep learning for a given problem.
Discuss how dataset size, feature types, nonlinear relationships, noise, interpretability, training cost, and inference latency affect the choice. Describe an evaluation strategy and implement a small comparison using a supplied tabular dataset, stating which model you would select and why.