University of Minnesota
Software Engineering Center

You are here

Your Laptop isn't Production: Engineering Your ML Models into Production

Date of Presentation: 
Wednesday, January 16, 2019
Do your ML products run on a data scientist’s laptop? Are they being run manually? Is the “production” version the first one that ran without crashing? Is it extremely challenging to update and redeploy your ML products? ML engineering and data engineering might not be the sexiest part of data science, but without understanding the engineering of building, deploying, optimizing, and monitoring ML systems, you’re handcuffing a model’s full potential. This talk will share some of the critical knowledge to engineering production-level machine learning products, instead of hacked-together systems. For the record, Donald thinks the engineering is the sexiest part of machine learning.