Final week, first day. From now, consolidate scattered knowledge into test-ready form, not learning new concepts. MLA-C01 spans four domains. Domain 1 (Data Preparation, 28%) and Domain 2 (Model Development, 26%) combine for over half. Master these two, you start exam with the majority. Today: data collection, feature engineering, Feature Store, Clarify combined with model training, tuning, evaluation in one flow.
Core message: Data preparation and model development connect via "reproducibility and consistency." Features used in training must match exactly in inference (avoid skew); data and hyperparameters in training must be tracked; bias and leakage detected at data stage.