Explain Feature, Model, and Validation Choices

You are interviewing for a Data Scientist role. Describe how you would approach an end-to-end machine learning project on large-scale data. In your answer, cover all of the following: - the standard data processing workflow from raw data extraction to modeling and deployment-ready outputs; - how you create and select features, including how you think about domain knowledge, missing data, feature leakage, multicollinearity, temporal stability, and feature importance; - which machine learning models you have used and how you decide which model to try first; - why you might choose Random Forest instead of XGBoost, and when XGBoost would likely be the better choice; - the differences between Logistic Regression and Random Forest in terms of assumptions, interpretability, nonlinearity, feature engineering needs, training behavior, and calibration; - how you validate model results, including train/validation/test strategy, cross-validation, time-based splits when relevant, metric selection for class imbalance, and how you check whether the model will generalize. Use a concrete project example if possible, and explain not only what you did, but why you made those choices.