About
AI Systems Builder & Trainer
The Builder & Educator
I build agentic AI systems and teach teams to run them on real operational workflows. Creator of Project Athena, an open-source local-first agentic memory system (590+ GitHub stars, 77 forks).
My engineering foundation is paired with formal adult pedagogy: I spent three years developing financial-advisory certification curricula for Singapore's financial sector and hold the WSQ Advanced Certificate in Training and Assessment (ACTA).
Background
Bachelor of Business Management (High Merit, Finance & Corp Comms), Singapore Management University 2011.
Passed Level I of the CFA Program, ChFC®/S & CLU®/S (conferred 2013), CFP® Modules 1–6 (2010), WSQ ACTA (2014)
Curriculum Development Officer at Singapore College of Insurance (SCI, 2012–2015) — MAS-regulated advisory tracks.
Hands-on AI Agent Workshops for Founders & Teams · Project Athena Agentic PKM Architecture
Philosophy: The Anti-Slop Protocol
I navigate reality as it IS, not as it SHOULD BE. In an era of AI-generated noise ("Slop"), signaling requires friction.
This isn't cynicism — it's operational clarity. Moralism is a luxury; physics is mandatory. The question isn't "Is this fair?" but "What are the actual forces at play, and how do I position myself to survive them?"
"Navigate, don't litigate."
The Bionic Operating Model
I operate as a bionic unit — human cognition augmented by AI at every layer.
Human Intent (Top 1% Nuance) + AI Scale (Bottom 99% Execution) = 1000% Leverage
This follows the Architectural Isomorphism principle: The User provides the Meta-Cognition (How to think), and the AI provides the Execution (What to do).
The AI isn't a replacement; it's a force multiplier. I provide strategic direction and high-nuance judgment. The AI handles research, execution, and documentation at scale.
AI Proficiency
I work with state-of-the-art frontier models daily — not as a casual user, but as a practitioner who understands prompt engineering, agentic workflows, and how to orchestrate AI systems for production-grade output.
Agentic Engineering Stack
- Agentic IDEs & Harnesses: Claude Code, Antigravity IDE, Cursor
- Protocols & Context: Model Context Protocol (MCP), dynamic skill injection
- Frontier Foundation Models: Claude 3.5 Sonnet / Opus, Gemini 2.5 Pro / Flash
- Engineering: Python, Supabase (pgvector), SQL, Git & GitHub Actions
AI Capabilities
- Prompt Engineering: System prompts, few-shot, chain-of-thought
- Agentic Development: Multi-step autonomous workflows
- RAG Systems: Vector search, semantic retrieval, cross-encoder reranking, RRF fusion
- AI Image/Video: Midjourney, DALL-E, Runway, Veo 2
Every project I ship is AI-native — from research and design to code and content. This isn't "using ChatGPT"; it's running a bionic production pipeline.
The Stack
Preferred Tools
- Agentic IDE: Google Antigravity
- Reasoning: Claude Fable 5, Claude Opus 5, Gemini 3.5 Pro
- Frontend: Astro (Static Site Generator)
- Backend: Python, Supabase (pgvector)
- Architecture: Flat-file Markdown, Protocol-Driven
Hard Passes
- Black-box RAG: If I can't audit the retrieval, I don't use it.
- Bloated Frameworks: No LangChain/AutoGPT. Primitives only.
- Platform-Locked Memory: My context stays in Markdown.
Current Project: Athena (v10.0.2)
Athena is an AI-Powered Digital Personal Assistant operating on the Zero-Point Codex architecture. It enforces the Triple-Lock Protocol (Search → Save → Speak) to ensure every output is grounded in context.
It's not a chatbot. It's a living architecture of protocols, frameworks, case studies, and calibrated insights that compounds over time. Every conversation becomes intellectual capital.
View the public repository →Want to work together?
Internal AI systems. Team AI training. Ops automation. Knowledge engines.
Get in touch →