Backend Developer at PT. Eigen Tri Mathema

In this role, I worked on building and improving the REST APIs for a Mini-ERP system covering Inventory, Company, POS, and Discount modules. Most of my time went into making sure the data flow was clean and reliable, from designing the schema to ensuring the operations stayed ACID-compliant and responsive even when traffic spiked. I also spent a good amount of effort fixing performance issues on both the backend and frontend. After cleaning up a few heavy queries, restructuring some request flows, and adding the right caching layers, we were able to cut latency by around 40%, which made the whole system feel much smoother. Part of the job involved maintaining older sync apps that handled online/offline data. These had a lot of edge cases and used to fail quite often, so I added idempotency handling, better conflict resolution, and a more predictable retry mechanism. This reduced sync failures significantly and made the system way more stable for users in the field. To support future growth, I also started introducing microservice-friendly patterns like DDD and event sourcing. Not a full migration, but enough to set a clear direction for scaling the platform in the long run.

Software Engineer at Ajobthing Sdn Bhd

I led the end-to-end migration of several high-traffic queries from SQL to Elasticsearch. My work included designing the indices, analyzers, and field mappings, as well as building the ingestion and backfill jobs needed to populate the new clusters. Before switching traffic over, I added parity tests to compare SQL and ES outputs so we could validate correctness and avoid regressions. The migration significantly reduced the load on our primary database, offloading roughly 40–60% of read queries. With Elasticsearch handling the bulk of the read operations, most queries became sub-second, which improved the overall responsiveness users experienced in the product. On top of Elasticsearch, I built REST/JSON endpoints that supported facets, filters, and pagination. I also created clearer API contracts and examples, which made integrator onboarding faster and more predictable, around 30% quicker based on team estimates. To ensure reliability after migration, I implemented a sample-based validation flow that compared SQL and ES results side-by-side, with an acceptance threshold to flag anomalies. I also collaborated with the team on relevance tuning. This involved adjusting query DSL, boosts, and analyzers, as well as building the instrumentation needed to support evaluation and annotation workflows and ongoing data-quality checks.