I build and operate AI systems that run in production: retrieval pipelines, evaluation + quality gates, and reliable cloud infrastructure for LLM-backed products. My work is backend-first, with a focus on measurable relevance, performance, and security.
🔎 Focus: AI platform infrastructure • Retrieval/RAG • Performance & security automation • Resilient deployments
📬 Contact: LinkedIn • Email
- 🧠 AI retrieval infrastructure: hybrid retrieval (graph + vector + structured filters), indexing, ranking, tuning
- ✅ Production readiness: load testing, latency/concurrency modeling, rollout safety checks
- 🔐 Security automation: OWASP ZAP scans wired into repeatable workflows
- 🛠️ Cloud reliability: active-active patterns, routing, TLS hardening, operational guardrails
- 🧑💻 Backend: highly adaptable across backend stacks (APIs, services, data pipelines, integrations). This is my primary strength.
- 🎨 Frontend: not my focus (I can integrate and support, but I don’t position myself as a frontend specialist).
Languages: Python • TypeScript/JavaScript • Bash
AI/Retrieval: LangChain • embeddings pipelines • hybrid ranking • evaluation workflows
Data/Stores: Neo4j • PostgreSQL/pgVector • TiDB • MySQL
Infra/Delivery: Docker • Nginx • Linux • active-active deployments
Testing/Quality: k6 • Locust • Playwright • OWASP ZAP
Most of my production work serves state government use cases, so the codebases are confidential. Some deployed products are publicly viewable:
- 🌐 Dayang chatbot (Sarawak services portal): https://service.sarawak.gov.my/web/
- ⚖️ Court-related project (public article reference): https://ekss-portal.kehakiman.gov.my/portals/web/home/article_view/0/5/1
- 📚 Malaysia public library chatbot (button-based): https://www.u-library.gov.my/portal/web/guest
These repos represent the kinds of systems I build (pipeline → retrieval → validation), even when production code is not public:
| Area | Repository | What it shows |
|---|---|---|
| 🎬 Local multi-agent AI app | agentic-video-analyst | offline inference + multi-agent orchestration + desktop app engineering |
| 🕸️ Graph ingestion + retrieval | neo4j-document-pipeline | graph modeling + retrieval API patterns for LLM workflows |
| 📈 Vector + hybrid experiments | tidb-vector-llm-testbed | relevance/scoring experiments, indexing tradeoffs |
| 🧬 Embedding pipeline | mysql-to-pgvector-embeddings | extraction → embeddings → pgVector semantic layer |
| 📚 Structured retrieval | faq-retrieval-system | structured query layer for grounded answers |
| 🧪 Performance testing | playwright-dayang, k6-for-custom-dify | UX + API load testing approaches for assistants |
| 🛡️ Security automation | zap-security-api | ZAP baseline/quick/full scan exposed via API |
| 🧩 Experiments | playwright-study, besu-ibft2.0 | targeted learning repos (testing + distributed systems) |
- 📏 Prefer measured improvements (evaluation + monitoring) over demo-only features
- ⏱️ Treat quality, latency, and security as release criteria
- 🔍 Build systems that are operable (clear failure modes, logs/metrics, runbooks)