01 / 05
Enterprise field operations2024 to now
ElxerOne
The platform more than 100 employees open before anything else, with location tracking that keeps working after the app is killed.
My role
Sole mobile engineer, architecture to release
- Enterprise
- Field operations
- Location
- Security
The problem
Elxer runs field operations across a city. Engineers on tickets, collections staff on the ground, attendance spread over multiple offices. Most of it lived in spreadsheets, phone calls and trust.
Trust does not scale. Neither does ringing someone to ask where they are.
What changed
Used every working day by more than 100 employees across field, collections and office teams.
Feature delivery got roughly 25% faster after the refactor, and the codebase became testable for the first time.
About ₹36,000 a year of SDK licensing avoided.
Attendance and commute claims turned into evidence instead of assertions.
What I did
5 moves
Built the whole platform in Flutter as the only mobile engineer on it. Attendance across multiple office locations, leave with approvals, ticket assignment based on live location, collections workflows, and automated commute tracking for fuel reimbursement.
Wrote persistent live location tracking that survives the app being terminated. Commercial SDKs that do this run about ₹36,000 a year. I built it in house instead.
Closed the door on location faking: developer options detection, mock location blocking, and device binding. A check-in now has to come from the right person, on the right phone, in the right place.
Replaced a slow third-party package with a Network Mapping Tool of my own. Image to location went from 26 seconds to 1 or 2.
Moved the core codebase to Clean Architecture and MVVM, added a feature flag system and set up CI/CD through Codemagic, so shipping stopped being an event.
Walkthrough
Screens
