Design a complete ML pipeline for a house price prediction system. Document each stage: (1) Problem definition, (2) Data collection sources, (3) Data cleaning & preprocessing, (4) Feature engineering, (5) Model selection (compare 3 algorithms), (6) Training & validation split, (7) Evaluation metrics, (8) Deployment considerations.