Project Overview
Nusantara Digital Sdn Bhd, a Cyberjaya-based social technology company, partnered with SCS Synergy to build Kawan, a community-driven social media platform designed specifically for Southeast Asian users. Unlike global platforms that prioritise individual content creation, Kawan centres around neighbourhood communities, local interest groups, and cultural events, reflecting the communal values deeply embedded in Malaysian and wider ASEAN societies.
Our team developed native iOS and Android applications using Flutter, chosen for its excellent performance characteristics and ability to deliver a truly native feel on both platforms from a single codebase. The application features a sophisticated content feed powered by a recommendation engine that prioritises geographically relevant content, local community discussions, and trending topics within the user's kampung or neighbourhood.
Kawan includes innovative features tailored to the region, including a marketplace for local sellers integrated with Malaysian e-wallet payment systems, a community events calendar with mosque and temple prayer time integration, and a multilingual interface supporting Bahasa Malaysia, English, Mandarin, and Tamil. The platform also features a unique "Gotong-Royong" feature that enables communities to organise volunteer activities and mutual aid initiatives.
We built the backend on a microservices architecture deployed on Google Cloud Platform, utilising Cloud Spanner for globally consistent data storage and Cloud CDN for low-latency content delivery across the ASEAN region. The real-time messaging system is powered by WebSockets with Redis pub/sub, supporting group chats of up to 500 members with instant message delivery.
The Challenge
The Southeast Asian social media landscape was saturated with global platforms that treated the region as a secondary market. Content moderation for multilingual posts mixing Bahasa Malaysia, English, and local dialects presented a significant technical challenge, as existing AI moderation tools performed poorly on code-switched text common in Malaysian online communication. The platform needed to handle content like "Jom makan lunch kat that new mamak" seamlessly.
User acquisition in a market dominated by established platforms required a compelling value proposition. Research conducted by Nusantara Digital showed that 65% of Malaysian social media users felt disconnected from their local communities online, with global platforms' algorithms favouring viral content over neighbourhood-level discussions. The technical architecture needed to support rapid scaling from launch to potentially millions of users while maintaining the intimate community feel that differentiated Kawan.
Our Solution
SCS Synergy designed a community-first architecture where geographic proximity and shared interests drive content distribution rather than pure engagement metrics. We implemented a custom content ranking algorithm that weighs local relevance, community participation, and content freshness, ensuring users see posts from their neighbours and local businesses alongside broader trending topics.
For content moderation, we developed a hybrid system combining a custom-trained multilingual NLP model with community-based moderation. The NLP model was trained on a dataset of 500,000 Malaysian social media posts spanning multiple languages and dialects, achieving 94% accuracy in detecting harmful content in code-switched text. Community moderators, elected by their neighbourhood groups, handle nuanced cultural context that AI cannot reliably assess.
The application was designed with a strong emphasis on data efficiency, critical for users in rural areas with limited mobile data. We implemented aggressive image compression, lazy loading, and a sophisticated offline caching system that allows users to browse previously loaded content and queue posts for upload when connectivity is restored. The messaging system uses binary protocols to minimise data consumption to approximately 40% less than competing platforms.