Executive Guide to Deploying, Monitoring, and Maintaining Models

Course details

With recent developments in the AI space, workflows for deploying, monitoring, and maintaining ML models have changed. In this course, Keith McCormick—an independent data miner, trainer, speaker, and author—breaks down the phases of an ML project and guides you through model evaluation, scoring, deployment, and model maintenance. Learn about data engineering and MLOps in the ML lifecycle, as well as the basics of ML modeling. Get a useful deployment checklist that you can use in model evaluation. Find out how to score traditional ML models, a “black box” model, and an ensemble. Go over batch and real-time scoring. Plus, explore model monitoring and the best frequency for model rebuilding.

Concepts

0. Introduction

01 - Getting models deployed

1. The Phases of a Machine Learning Project

02 - Data and supervised machine learning
03 - Data engineering and MLOps in the ML lifecycle
04 - Why ML projects fail to be deployed
05 - The basics of ML modeling

2. Model Evaluation

06 - The business evaluation phase
07 - A deployment checklist

3. Scoring

08 - Scoring traditional ML models
09 - Scoring a black box model
10 - Scoring an ensemble

4. Deployment

11 - Batch vs. real-time scoring
12 - Data prep and scoring
13 - Combining batch and real-time scoring

5. Monitoring and Maintenance

14 - What is model monitoring
15 - How often should you rebuild

Conclusion

16 - Next steps