Developer Productivity at Scale
summary
Toolchain speed and reliability drive developer behavior: teams with 4‑minute vs 1‑minute builds ask for feedback far less; reducing a 1‑minute build to 0.6 minutes yields 44 days saved for a team running ~1,000 builds/week. Larger orgs cutting 9→5 minute CI save ~5,200 days (25% R&D). Concrete fixes include build caching (example: a 300k‑LOC, 100‑module Maven build 20→<2 minutes) and distributed test execution; capture per‑build data to detect regressions, flaky tests and misconfigurations (128→1 thread caused by a dotfile).
AI-written summary · about
chapters (15)
- 0:00Introduction to talk
- 0:42Developer productivity importance
- 2:50Impact of build times on feedback
- 9:03Consequences of long builds
- 13:00Case study on build time reduction
- 16:34Build caching explained
- 22:15Benefits of build caching
- 25:01Modularization and monoliths
- 26:28CI load reduction with caching
- 27:55Data-driven build improvements
- 31:08Why build data matters
- 33:15Troubleshooting build issues
- 36:28Reactive vs proactive fixes
- 39:26Flaky tests challenges
- 43:34Summary and resources
related talks
36mDeveloper Productivity Engineering: What's in it for me?
02h 31mDeveloper Productivity Engineering - Extended Session
45mDeveloper Productivity Engineering – The Next Big Thing in Software Development
47mHow to win in software development with tighter feedback loops
27mThe Need for Speed (in your CI/CD setup)
54mDeveloper Productivity Engineering Deep Dive
27mDeveloper Experience: It’s Time to Start Complaining
44mDeveloper Productivity Engineering: What's in it for Java Developers?