Model Deployment and Monitoring

This unit teaches students how to prototype interactive model demos, create publication-quality visualizations, deploy models as APIs, and establish basic monitoring for both traditional models and LLMs. It builds on prior model development skills and prepares learners for production hardening, observability, and scaling in the final unit.

Learning Objectives

• Demonstrate how to prototype interactive ML interfaces using Gradio, producing reproducible demos that map model inputs to outputs with appropriate UI controls
• Create and interpret visualizations with Plotnine to explore model predictions, feature importance, and evaluation metrics for clear communication of results
• Deploy machine learning models (including a simple Random Forest and LLM wrappers) as RESTful APIs, documenting endpoints, dependencies, and basic containerization steps
• Design and implement basic monitoring and logging pipelines to detect model performance degradation, data drift, and latency issues, and evaluate mitigation strategies