Your question is Feature Engineering for Noisy Data. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You are working with a supervised learning dataset that has many noisy, correlated, and partly missing features. Several candidate models have been unstable across validation splits, and the team wants a cleaner approach to feature construction and selection.
How do you approach feature engineering for noisy, high-dimensional datasets?