Kelola Group pushed my work into nationwide operations. The company ran across Indonesia's major fishing and seafood regions, and a lot of field reporting still depended on WhatsApp messages and email threads. The gap between what the business needed and what the tooling provided was obvious.

I built an Android app in Kotlin with a PHP backend so regional sales reps could capture field data in a structured way. That work mattered because the app was not just another reporting surface. It replaced fragmented manual processes with something management could actually trust and use.

The field context shaped a lot of the implementation decisions. Data had to make sense across regions, the UI had to stay simple enough for fast use on the ground, and the backend had to support workflows that were already messy before software entered the picture. It was a good lesson in how "simple product" and "simple business process" are rarely the same.

I also built an e-commerce platform for the company's direct sales arm, using Dolibarr and Vue.js. That made the role a full stack balancing act: mobile flows in one product, web commerce in another, and business logic that needed to survive in both places.

Kelola taught me how different systems inside one company can still depend on the same engineering habit: translate ambiguous operational needs into concrete product behaviour. Whether the user was a field rep or an internal commerce team, the work was the same at its core.

That period widened my range. I was still comfortable going deep on implementation, but I had started to see more clearly how software earns its keep inside a business: fewer manual steps, cleaner inputs, better decisions, and less operational drift.