Lead Android
engineer

What I inherited

A consumer calling app on Google Play that people use to stay in touch with family. When I took over the Android side it sat at 2.4 stars, with around 20 recurring crash and ANR issues on an ageing XML codebase. People weren’t leaving for missing features — they were leaving because the app failed mid-call. The reviews also pointed at confusing UI and buttons that looked wrong in certain states.

What I did

Instead of chasing individual reports, I worked from causes.

• Clustered crashes by root cause — most traced back to memory leaks and to library callbacks returning null with nothing to handle them, so twenty separate-looking issues collapsed into far fewer fixes.
• Treated the review backlog as a defect list, not feedback, and prioritised each problem by how many people it affected — design and state bugs included.
• Moved to staged rollouts, so anything that slipped through was caught at a small percentage of users instead of everyone.
• Grew the app from audio-only calling to audio and video.

What I learned

When a product is struggling, the instinct is to ship new features. It’s usually the wrong one. Fixing what already hurts people earned back more trust than anything new would have.