Payam Zahedi
Software engineer
Flaconi
I build mobile apps for a living and make them prettier, funnier and faster for fun. I'm a software engineer with 8+ years in mobile and the author of toastification, an open-source Flutter package that makes toast notifications beautiful and effortless. I also lead the Persian Flutter community, where I help developers learn, share and grow together. Based in Berlin, I'm on a mission to prove that "efficient" and "delightful" belong in the same widget.
From Review Bottleneck to Auto-Merge: AI Code Review for Flutter Teams
Session
AI coding tools helped our team produce Flutter code faster, but pull requests started waiting longer for review. The bottleneck had moved from implementation to the PR queue.
In this talk, I'll show how we built an AI reviewer for our production Flutter app with GitHub Actions. I'll compare Copilot CLI, Cursor CLI, Claude routines, and Claude Code, explain why we chose our current setup, and show how it helps us review changes faster.
You'll see how we moved beyond generic comments by teaching the reviewer our Flutter architecture, conventions, best practices, Effective Dart rules, and team standards. I'll also show how a code graph gives it the context to review a change beyond the diff while reducing noise and token usage.
You'll leave knowing:
* Why AI coding moves the bottleneck to review
* How to teach it your Flutter and Dart practices
* How code graphs improve context and review quality
* Which PRs can be auto-approved, and when a human should stay involved