Experience
Where I have worked
Software Engineer — Defence Science and Technology Agency
- Led the technical design of a data archival initiative: proposed a self-hosted lakehouse architecture, aligned it with DSTA-wide data standards, and presented the design to leadership for approval.
- Prototyped the lakehouse end to end on Kubernetes toward an air-gapped deployment — Spark, Trino, Iceberg + Nessie, MinIO, NiFi, Elasticsearch — secured with Keycloak, OPA, and Istio, with a self-contained supply chain via Harbor and GitLab.
- Built and operate an event-driven job-scheduling service in production (Spring Boot, Kafka, Quartz, Redis on Kubernetes), designed for 100k+ events/hour.
- Fine-tuned XTTS v2 on 2,171 hours of Singaporean speech (voice similarity +5–10%), with a Streamlit + FastAPI demo and MongoDB embedding caching for real-time voice cloning. A fine-tuned XTTS v2 voice also narrates AI Deck Studio.
Data Engineer Intern — Razer Inc.
- Led prototyping of an AI-powered unit-test generator using RAG (LangChain) and Pinecone for contextual code retrieval and automated test generation. Read more.
- Refactored ETL from Pandas to optimized SQL on AWS Redshift and migrated materialized views to dbt, improving ingestion stability and data freshness for the recommendation engine. Read more.
Full Stack Developer Intern — IBM Singapore
- Built an ML-powered document search on AWS Kendra — a custom S3 data source connector, type-ahead suggestions, and relevance tuning — plus Spring Boot REST services behind a backend-for-frontend. Read more.
- Shipped a recently-viewed-documents feature spanning services across VPCs, and raised code coverage 60%+ with Mockito-based tests to improve release confidence.
Research Intern — A*STAR
- Researched time-series forecasting of intermittent supply-chain data using LSTM/GRU models, presenting findings and Python visualizations to stakeholders. See the poster.