Predictive Test Selection

For most projects, a majority of the software build time is spent
running tests. Often, many of the tests that run in a single build are functionally unnecessary.

Despite this, common test impact analysis methods will run them out of an abundance of caution. While safe, this practice introduces unnecessary delays to the developer feedback cycle. To mitigate this, Gradle Enterprise is proud to announce Predictive Test Selection, which significantly reduces testing time by leveraging machine learning to avoid irrelevant tests. Inspired by Predictive Test Selection at Meta, this system proactively develops a test selection strategy by learning from historical changes and test outcomes from your Build Scan data.

In this webinar, you will see Predictive Test Selection in action, learn how it works, and participate in a live Q&A with the lead data scientist, Eric Wendelin.