Apps at Scale: An Industry Discussion
summary
Panel compares real engineering choices: monorepos vs many repos, build systems and tradeoffs (Gradle + caching, OkBuck, Buck and Bazel, ongoing move to Bazel), concrete scale metrics like 3500 modules and 1200 micro‑modules, Airbnb’s “light” app that cuts builds ~75% for local work, Kotlin adoption stuck at ~10–20% due to compiler and build pain, weekly release trains, heavy use of static analysis, feature‑flag patterns and when experiments actually roll back vs ship. Practical tooling pros and cons.
AI-written summary · about
chapters (19)
- 0:00Introduction and panel setup
- 1:14Panelist introductions
- 2:36Team structures at scale
- 5:45Developing for large teams
- 8:13Mindset shift at scale
- 10:43Day-to-day in platform teams
- 13:33Prioritizing platform work
- 15:43Build system challenges
- 23:08Staying close to mainline tools
- 23:28Communication at scale
- 25:34Codebase structure choices
- 31:13Monorepo benefits and trade-offs
- 36:08Build system comparisons
- 44:07Kotlin adoption strategies
- 51:29Feature flags and experimentation
- 59:37Release process and build trains
- 1:05:11Engineering culture insights
- 1:13:39Code review and ownership
- 1:24:35Top tips for scaling teams
related talks
01h 00mAndroid at Scale
40mGrowing Pains: Scaling and Re-architecting Apps Under Fire
43mThe Engineering Playbook for Growing Android Teams
41mBuilding Mobile Apps At Scale
37mA Complete Journey of Android Modularisation
40mModularizing Android Applications
23mModularizing Your App
34mIntro to App Modularization