AI product engineer building frontier AI systems for education.
I work across agent behavior, memory, retrieval, evals, backend infrastructure, and product interfaces.
At Pandai, I’m building P&AI, an open-source tutoring agent for students in Malaysia and Brunei. I like early-stage startups where I can listen closely to users, turn rough ideas into working products, and improve them through real feedback.
- 🤖 P&AI — An open-source tutoring agent for curriculum-based practice
- 🖥️ Ernie — An experimental macOS workspace that keeps Prime Agent sessions with their repositories and Git worktrees
- 🌱 Garden — Run Eve on your own infrastructure
- 🦊 Agent Lite — A compact
agent-browserfork built around Camofox and Camoufox-powered Firefox - 🧰 Skills — Reusable instructions for AI agents and the people who work with them
- 🚢 Ship LLM — Typed, streaming LLM calls through Python decorators (archived)
- 📈 Capstone — Property valuation tools built as executable Marimo notebooks
- 🔎 Retrieval demo — An interface for inspecting retrieval results and their sources
- 🧪 Simkit — A simulation app with a focused Svelte interface
- 🧠 RLM #40: persistent multi-turn environments — Introduced a runtime-checkable
SupportsPersistenceprotocol with versioned state history, lifecycle cleanup, and regression tests
- In defense of Electron — The desktop software Electron made possible and what its successors should preserve
- The best software is yet to be made — Ernie and task-shaped interfaces for recursive agents
- Enduring agents, disappearing machines — Separating the lifetime of an agent from the machines it uses
- I make music and software — Stage fright, listening as craft, and learning to make before feeling worthy
- Building curriculum-grounded AI tutors
- Engineering agent memory, retrieval, evals, and safety controls
- Taking early-stage products from rough ideas to reliable systems