Building Mobile Apps At Scale
summary
Defines two concrete scales: 100–200 Android engineers in a mono repo versus billions of daily active users, and measures scale by DAU/WAU and device diversity. Recommends specific tooling and practices: Firebase Remote Config feature flags and gradual experiments, Crashlytics and heavy logging for rare-device failures, CI/CD with per-commit tests, app distribution for QA, holiday code freezes, and automated rollouts. Describes real tradeoffs—slow library upgrades, long SDK bumps, and using kill switches to stop crashes immediately.
AI-written summary · about
chapters (10)
- 0:00Introduction and speaker background
- 1:54Defining scale in mobile apps
- 7:51Active users and seasonal spikes
- 10:41Device diversity challenges
- 15:13Two types of scaling problems
- 15:54Feature flags for safety
- 22:58Experimentation frameworks
- 27:15Managing many developers
- 37:06Handling long-tail issues
- 40:15Conclusion and key takeaways
related talks
01h 00mAndroid at Scale
35mRelease Android apps globally with no headache and minimum stress
37mTackling Android Development at Scale
43mThe Engineering Playbook for Growing Android Teams
01h 28mApps at Scale: An Industry Discussion
48mMobile Developer Productivity at Scale
41mMobile DevOps
10mWelcome to Mobile @ Scale