
Launchable and Test Intelligence

#40 - Data-Driven DevOps With Launchable - Kohsuke Kawaguchi
Topic 6 of 11
Published May 24, 2021 · Clip 21:14–37:13 (15:59)
Kohsuke explains Launchable: a test recommendation engine that uses machine learning to predict which subset of tests to run for given changes and in what order, to catch failures faster. He describes how it integrates with minimal intrusion via a small CLI, records build materials and test results, and improves predictions over time. He shares customer stories, including a Bazel user avoiding a two-year, four-engineer build migration and a two-person team halving a 20-minute test cycle, plus his vision of a self-tuning delivery pipeline driven by data.