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