The problem

A property app that has been live and growing for years accumulates weight: Java screens written before Kotlin, ButterKnife bindings the ecosystem has moved on from, endpoints pointing at an older gateway, and crash reports that arrive faster than they can be triaged. Rewriting is not an option when the app is in daily use.

The approach

Contributed to an incremental migration against the live codebase. Java was converted to Kotlin and ButterKnife replaced with ViewBinding screen by screen, so releases did not depend on the whole migration finishing. Edge-to-edge support was added for Android 15+, endpoints were moved onto the new API gateway, and FreeRasp was added for runtime hardening. I also handled production triage through Crashlytics and shipped fixes via GitLab CI and Fastlane.

Brighton Real Estate runs a nationwide property marketplace covering listings across Indonesia, backed by thousands of agents in dozens of cities. The Android app was already live and mature when I joined in November 2024 — built on MVP with Dagger 2, with a long tail of Java and ButterKnife still in place. I contributed to the existing production codebase rather than a greenfield build. My work included incremental Java → Kotlin migration, replacing ButterKnife with ViewBinding screen by screen, adding edge-to-edge support for Android 15+, migrating endpoints onto a new API gateway, and hardening the app with FreeRasp. I also handled production triage: reading Crashlytics, reproducing crashes and NPEs against real user reports, and shipping fixes through GitLab CI and Fastlane.