Geolocation-based social app for iOS
A social app with over 3 million users, built by cross-functional teams that decided with data. I worked on it for almost nine years, three of them as iOS chapter lead.
Context
The app has more than 3 million users, 5 million at its peak, mostly in the DACH and MENA regions. I joined in January 2018 and stayed until October 2026.
My role
I was an iOS chapter lead for around three years, and an engineer on the product the rest of the time. The teams were multi-functional. We tried Spotify-style squads and later Shape Up.
What I worked on
- Channels and real-time chat, built from scratch.
- Many iterations of features that were already running.
- Feed performance improvements.
- Small experiments in feeds and post details.
- A redesign and the introduction of dark mode.
How we worked
- Data first. Decisions came from user interviews and tracking, not opinions.
- Experiments. Many A/B tests. Experiments sat behind feature flags and we measured engagement in selected cities.
- Stability. We aimed for 99% crash-free sessions while mostly shipping new features.
Codebase
It started as 100% Objective-C and was 99% Swift by 2026. The architecture was MVVM with routers, data lived in Realm, and releases ran through fastlane and Bitrise CI.