You are working on a supervised learning model and need to improve its performance without making it unstable or hard to maintain.
What is your experience with machine learning model optimization techniques?