Skip to main content
Professional

Kelas Pintar — Scalable Learning Management System Platform

Kelas Pintar · Mobile Developer

Cross-platform LMS mobile application serving large-scale digital education experiences for students and educators across Android and iOS, with a focus on frontend architecture modernization, reusable component systems, and automated regression validation.

Kelas Pintar — Scalable Learning Management System Platform

Overview

Kelas Pintar is a production-grade Learning Management System platform delivering accessible, scalable digital education experiences to students and educators across Indonesia. The mobile application supports interactive learning workflows, educational content delivery, and modern mobile-first accessibility — serving a large user base across Android and iOS. My contribution focused on frontend architecture modernization, building reusable UI systems, improving runtime stability, and deploying confidently through automated regression validation.

Problem

As the platform scaled to serve more users and educational institutions, the mobile codebase began showing signs of architectural strain — inconsistent component patterns, weak TypeScript coverage, and a manual QA process that slowed release confidence. Feature delivery was bottlenecked by the absence of shared UI primitives, and production regressions were difficult to catch before they reached users.

Solution

I contributed to a systematic modernization of the frontend architecture — introducing strict TypeScript across new and refactored modules, extracting shared component systems that could be composed across features, and integrating Katalon Studio for automated regression validation. The result was a codebase that scaled with the team rather than against it: new features could be built faster, regressions were caught earlier, and deployments to both app stores became a lower-risk operation.

Architecture

The application is built on React Native targeting Android and iOS from a shared codebase. MobX manages application state with observable/computed patterns well-suited to the reactive data flows of an LMS — live lesson status, progress tracking, notification states. REST API integration connects to the Kelas Pintar backend. Katalon Studio sits outside the app itself as the regression automation layer, running end-to-end scenarios against production builds before store submission. App Store Connect and Google Play Console manage the release pipeline.

Lessons learned

Working on a large-scale educational platform reinforced how much architecture decisions compound over time. Strict TypeScript from the start would have prevented entire categories of runtime bugs we encountered during modernization. Automated regression testing is most valuable when it covers the happy paths that users hit most often — not exotic edge cases. And in cross-functional teams, the quality of a feature is as much a function of engineering-product-QA alignment as it is of code quality.