Flutter at scale
summary
BMW adopts Flutter as a single codebase for iOS, Android and web supporting multiple brands and regional flavors, using trunk‑based development with scheduled train releases and feature flags to flip capabilities per market. Core tooling includes a modular architecture, the flutter_bloc state pattern, centralized logging via lumberdash clients, and CI/CD in Docker that enforces 100% test coverage plus thousands of unit/widget/integration tests and performance gating with Oussie screenshots and 60fps thresholds.
AI-written summary · about
chapters (23)
- 0:00Introduction to Flutter at scale
- 1:58Defining goals and philosophy
- 4:27Regular releases with train model
- 6:52Cloud-centric mobile approach
- 7:35Developer friendly practices
- 8:37Developer scalable architecture
- 10:04Performance and crash metrics
- 11:08Safe experimentation with feature flags
- 12:30Continuous deployment strategy
- 14:15Learning from past experiences
- 15:40High-level architecture overview
- 16:22Vertical module structure
- 18:32FlutterBlock state management
- 21:18FlutterBlock implementation
- 23:05UI integration with FlutterBlock
- 25:11Logging with Lumberdash
- 27:36Analytics and remote debugging
- 28:37Automated testing and quality
- 30:01CI/CD pipeline setup
- 31:04Performance testing with OSSie
- 33:54Key takeaways and recap
- 35:40Developer experience importance
- 36:45Closing and resources
related talks
42mOur Journey to Flutter
45mFlutter: A journey of productivity and happiness
22mIntroduction to Flutter
32m"Flutter: How we're building a UI framework for tomorrow at Google"
46mAdopting Flutter at Scale
36mFlutter Add-to-App: The Good, the Bad and the Ugly
09mFlutter for Enterprise: How your company should be building with Flutter at scale
41mWhy is Flutter the mobile SDK developers deserve...