100 Million Lines of Code, Millions of $ Saved, 0 Wasted Builds; AI Powered CI @ Atlassian Scale
summary
Atlassian describes a CI overhaul that cuts build time 80% and saves over $4M/year by using predictive test selection that skips 50–95% of tests per change, a flake management system that auto-skips identified flaky tests and surfaces alerts, and a dependency graph plus predictive orchestration to run only affected builds. They migrate 5,000 Playwright tests with 60% automated, fix 6,000 accessibility bugs at 90% AI acceptance, and report 40% better reliability.
AI-written summary · about
chapters (10)
- 0:00Introduction and context
- 1:07Scaling challenges at Atlassian
- 4:36Why improve developer productivity
- 7:02Monorepo scale and complexity
- 9:09Key infrastructure solutions
- 13:25Results and impact metrics
- 14:50AI for repetitive tasks
- 15:51Next generation CI with AI
- 17:38AI-powered developer workflow demo
- 24:54Q&A and closing
related talks
44mSimplifying and Scaling Software Delivery
24mDevelopers and AI Enemy or Frenemy Leveraging AI and Atlassian to empower developers
06mTriple Threat: AI Technologies Reduce Testing Costs
58mDeveloper Joy – How great teams get s%*t done
21mAI-driven optimization of CI/CD pipelines
35mBeyond the chatbot, beyond the IDE, where can AI help your CI/CD
39mIt's Not Called Continuous Integration (CI) for Nothing
26mAccelerating DevOps with AI