You're comparing several candidate models for a supervised learning problem and need to improve generalization without overfitting. You want to explain how you optimize models in practice, from selecting the right complexity to tuning training settings.
What is your experience with machine learning model optimization techniques?