Developed a fraud detection model using Python, XGBoost, and SMOTE on an imbalanced Kaggle dataset (284K+ records, 492 frauds). Achieved 95%+ recall by performing EDA, applying threshold tuning, and ...