Applied Model Development and Evaluation

This capstone unit integrates practical model training, evaluation, and deployment workflows with emphasis on reproducibility, optimization, and responsible practice. Students will execute and optimize experiments, perform feature engineering and data verification, blend and validate models, and assess ethical and decision-making consequences to prepare for advanced study or workplace application.

Learning Objectives

• analyze model performance, estimation variance, and data leakage using cross-validation and diagnostic techniques to select, validate, and ensemble models (blending and stacking)
• apply feature engineering, feature selection, data verification, and dataset augmentation strategies to detect outliers, verify sources, and improve model generalization
• demonstrate reproducible and efficient execution of ML code and notebooks by using hardware acceleration, executing pipelines, sharing artifacts, and troubleshooting runtime issues
• evaluate ethical, legal, and societal implications of machine learning solutions by identifying discriminatory practices, balancing automated decision-making, documenting consequences, and outlining mitigation and help-seeking procedures