Florence: Production Mobile Software at 250k+ Users

TL;DR

Florence connects nurses and carers with shifts at care providers across the UK. Software that real people's work schedules depend on. As a Full Stack Developer from 2022 to October 2025, I owned the iOS and Android release cycles for a platform with 250,000+ active users: feature development, deployment, and the monitoring that catches problems before users do. It's the third leg of how I work: the same person who designs the mechanism and writes the firmware also ships and operates production software at scale.

The context

A shift-booking platform is unforgiving software: a bad release doesn't just annoy users, it interferes with people getting paid and shifts getting filled. Release discipline (staged rollouts, monitoring, the judgment call of when to hold and when to ship) was the core of the job, across both the workers' app and the business side of the platform.

What owning a release cycle actually means

Feature work is the visible part; the ownership is everything around it: taking a change from development through review, staged release to two app stores, and post-release monitoring, then deciding what the data says about the next release. Doing this at 250k+ active users for three and a half years is where my production-software discipline comes from, and it transfers directly to robot fleets: telemetry, staged rollouts, and safe failure behaviour are the same problems in different clothes.

Results

• 250,000+ active users on the platform whose iOS and Android release cycles I owned
• Three and a half years of continuous production releases across two app stores

Gallery