Three months. That's how long it took us to realize we'd made a mistake.
Not because anything obviously broke. The screens rendered. The navigation worked. The API calls went through. From the outside, progress looked normal — features were getting built, PRs were getting merged, the project was moving.
What we found when we actually dug in was code that worked in the way a demo works. Functional under controlled conditions. Quietly falling apart under real ones. Performance issues on mid-range Android devices that nobody on the team had tested on. Platform-specific bugs that had been sitting in the codebase for weeks, invisible because everyone was developing on iPhones. A FlatList implementation that would have caused serious problems the moment a real user with real data volume opened that screen.
The developer wasn't bad at JavaScript. He was actually pretty good at JavaScript.…