How Your ML Model Will Fail - And How to Prepare For It
Explore the critical aspects of maintaining machine learning models in production with this insightful 27-minute talk from the Toronto Machine Learning Series. Emeli Dral, CTO and Co-founder of Evidently AI, delves into the unique challenges of ML system failures and how to address them. Learn about the various ways models can degrade and break in production, and discover effective strategies for analyzing performance, detecting data drift, and monitoring data quality. Gain practical knowledge on setting up a pragmatic monitoring strategy to ensure the longevity and reliability of your ML models. This presentation is essential for data scientists and ML practitioners looking to enhance their skills in model maintenance and troubleshooting beyond the initial deployment phase.