Hong Kong · HKUST Economics · Founder
Kun Lap Ming.
Known as MM
An economics student who builds with AI. I teach AI in plain words, I'm moving into AI training, and I publish the tools I make for my own week. Outside class, I'm teaching myself finance.
ScrollThree things, one thread.
-
SEP 2026 –
Phantom UniversityCofounder
Practical AI, taught in short English Reels by an AI avatar host we built and run.
-
2026 –
AI trainingFocus
Writing and reviewing the prompts and answers that AI models learn from. It's the work I'm moving into next, and it starts from how I already test my own agents.
-
2026 –
Open source AI tools8 public repositories
Skills and agents that make an AI ask better questions, check its own answers and plan before it acts. Each one started as a problem in my own week.
Selected repositories.
- 01 · SKILL Rough Idea to Pro PromptTurns a rough request into a production ready prompt, waits for approval, then runs it. Works across AI models.
- 02 · SKILL AI Answer Fact CheckerLists the claims in an AI answer that could be wrong, rates each one's risk, and says exactly where to check it.
- 03 · SKILL Vague Goal to Clear PlanTurns a vague wish into a scoped brief with a definition of done, after at most five short questions.
- 04 · SKILL Out of the Box Idea GeneratorLists the obvious answers first, breaks the hidden assumptions in a problem, then returns ideas that each come with a cheap test.
- 05 · SKILL AI Daily Accountability SecretaryReport your day, get ON TRACK, MIXED or DRIFTED against your own goals, a seven day scoreboard and one nudge for tomorrow.
- 06 · AGENT · PYTHON Investing Agent“Is it a good time to buy?” answered as a timing checklist for stocks, ETFs and crypto. Conditions, never a verdict.
- 07 · PROMPT Resume BuilderOne prompt for any AI that scans a resume, asks a few questions and rebuilds it as one clean page. Never invents facts.
- 08 · MACOS APP · SWIFT Mac Desktop Productivity WidgetA free macOS widget: news, calendar reminders, an hour tracker and a quick copy vault. No AI, no account, no tracking.
How I work.
I don't come from computer science. I build by directing AI: I write the brief, set the rules, test what comes back and keep only what works.
Brief
Say exactly what done looks like before anything gets built.
Rules
Hard limits the model must keep: what it never claims, never sends, never invents.
Test
Run it on my real week and check every output against the brief.
Publish
Write it up plainly so anyone can use it, in any AI.
Study & credentials.
BBA in Economics
The Hong Kong University of Science and Technology
AI Fluency: Framework & Foundations
Sep 2026
Introduction to Generative AI
Sep 2026
Find me here.
Let's talk.