Senior Product Manager · Native AI Builder
Senior PM with ~10 years across EdTech and Fintech. Currently shipping a portfolio of production native AI products on top of Claude + MCP — multi-workspace vector RAG (42K+ sources), eval frameworks, knowledge audit, ADHD-first inbox triage, AI-first developer tools. I write what I learn at Vuihoc.AI.
Shipping production native AI products: multi-workspace vector RAG (42K+ sources), LLM eval frameworks, knowledge audit pipelines, MCP servers, OAuth flows. Foundation patterns reused across 10 shipped projects.
Senior PM ~10 years across EdTech and Fintech. CSPO + PSM I. Translating ambiguous priorities into structured programs with measurable business impact and clear executive accountability.
Driving alignment across engineering, design, finance, customer success, and partner organizations. Comfortable in deep technical reviews, customer onboarding workshops, and exec readouts.
Sharing what I learn — research notes on LLMs, MCP, RAG patterns, infra setups. Open notebook for AI/ML, prompt engineering, and product engineering experiments.
Senior Product Manager with ~10 years in product. By day I work on B2B SaaS for early-childhood centres in Singapore. By night I ship production native AI products on Claude + MCP — currently 10 shipped: a multi-workspace personal RAG (42K+ sources), an LLM eval framework, a cross-source knowledge audit, an ADHD-first inbox triage agent, an AI-first text-to-diagram DSL, a Force-Click Mac translator, an AI vendor-doc crawler, a voice assistant, an ADHD sleep coach, and an internal PM proposal.
A cross-section of AI infrastructure, vertical apps, and writing — from idea to shipped in days, not quarters.
A production-ready personal RAG system over 42,000+ knowledge documents — multi-workspace semantic search (LL work, MindX consulting, personal, shared research, AI vendor canon) accessible from Claude Desktop, Claude.ai web, and iOS app via the Model Context Protocol.
A personal health coach over Garmin data lake with ADHD-friendly nudges. Daily morning ping + Saturday digest + Calendar pre-meeting breath alerts via Telegram. Goal — weekly avg sleep score ≥80.
Voice-first agentic assistant running native on macOS — push-to-talk or continuous mode, controls 40+ Mac actions (open apps, type text, click, browser tabs), searches personal KB, multi-LLM (Cloud Haiku / Local Hermes / Apple Foundation Models). Pure Swift, ~50MB DMG, no Electron.
A local-first AI inbox triage layer that classifies threads P0/P1/P2/NOISE with cross-channel verification — cutting daily triage from 90 minutes to 8 minutes.
A text-to-diagram DSL whose primary user is an LLM, not a human — token-efficient syntax, deterministic SVG layout, and validator errors that LLMs can repair in one shot.
Before building 'Jira for AI agents,' I ran one Claude prompt that fanned out into 111 sub-agent runs — roughly twenty minutes and 3.6 million tokens of work — across five research angles searched in parallel, 29 sources pulled, 133 claims extracted, then three independent skeptics voted on every claim. Only 9 survived. The verdict killed the idea in minutes. A PM's field notes on using AI to disprove yourself, cheaply.
May 28, 2026I ran two AI coding agents — Claude Code and Antigravity — through 100 turns of pair-programming over a Redis Streams channel I built, while they wrote real code against a live Redis. The code wasn't the lesson. What kept them working together was the harness. Field notes on engineering the scaffolding around an LLM.
May 24, 2026Personal-RAG kept hallucinating vendor specs because Marc's notes and Anthropic's docs lived in the same index. The fix: a dedicated canon workspace, a crawler that only ingests authoritative sources, and a routing rule that prefers canon when the question is about pricing, model size, or API behavior.
Open to collaborations, speaking, or just to say hi. Find me on the platforms below or shoot me a message.