Senior Full Stack Engineer with 10 years of experience — from React & Node.js backends to agentic AI pipelines. Currently exploring the frontier of LLM-powered applications with LangChain, Claude API, and MCP.
I'm a Singapore-based Full Stack Engineer with a decade of experience building and shipping production web applications — from early-stage startups to enterprise cloud platforms at Oracle.
My work spans React & TypeScript frontends, Node.js/Express backends, CI/CD automation, cloud infrastructure on OCI, and — most recently — agentic AI systems using LangChain, Claude API, OpenAI, and the Model Context Protocol (MCP).
I've led frontend teams, mentored engineers, and owned end-to-end delivery on complex cross-functional projects. I'm energised by the intersection of solid engineering fundamentals and what's becoming possible with LLMs.
Led end-to-end full stack delivery for Java Management Service on OCI — from React/TypeScript frontend architecture to Node.js/Express backend services, CI/CD automation, and production releases. Built tooling that eliminated manual QA overhead and automated security alerting for enterprise Java deployments.
Led frontend development for a mission-critical mining safety application deployed globally by Rio Tinto. Built mobile-optimised, offline-resilient React UI consuming REST and GraphQL APIs in low-connectivity field environments.
Built and maintained a customer-facing e-commerce platform end-to-end — designing REST API endpoints, configuring servers, managing databases, and building Vue.js frontend components for product catalogue, cart, checkout, and POS.
Mentored students through a 12-week intensive full stack bootcamp, conducting code reviews, pair programming, and 1-on-1 guidance across React, Node.js, Express, PostgreSQL, and MongoDB.
A LangChain agent pipeline that ingests startup pitch decks (PDF), extracts structured data across key investment dimensions — team, market, traction, financials, competitive moat — and outputs a scored metrics report for VC analysts.
Currently exploring ideas at the intersection of agentic AI and real-world operational tooling. Check back soon or get in touch if you want to build something together.
The role is evolving fast with AI tooling. Here's what I think it looks like now.
PDF parsing is messier than you think. Here's the stack that actually held up.
After months of building with raw APIs, MCP shifted something in how I architect agents.
Blog launching soon — follow on LinkedIn for updates
Open to forward deployed engineering, full stack, and agentic AI roles in Singapore. Also happy to talk shop about LLMs, agent design, or anything in between.