Most AI conversations happen in decks. Mine happen in production.

I'm YY — COO of Laimi, building AI infrastructure for Asia's travel loyalty market. I run agent systems across multiple ventures and document what actually happens when AI meets real operations — not the theory, the friction.

YY Sng working on a laptop

What I Build

Agent systems. Running in production.

Evergreen

live

Not a website you update. A website your agent manages. Building and maintaining personal brand sites for senior operators, coaches, and consultants on the Evergreen architecture.

AI OS Brain

live

A personal AI brain built on Supabase with pgvector — stores insights, priorities, and context across all ventures. Every Claude session loads from it.

BD Research Agent

live

An agent that researches BD leads, classifies fit, and saves structured intel to Supabase automatically. Powers Laimi partner outreach.

Personal Content OS

live

The system behind my LinkedIn content — tracks performance, runs weekly reviews, generates new posts grounded in what worked.

Not a website. A system.

Evergreen is one place that holds what you do, and keeps up as it changes. Most people's published record sits still while their work moves — this keeps the record alive instead. It's the productized version of the system running this page.

Where I Operate

Three ventures. One operating system.

  • Laimi → AI-native travel infrastructure for Asia's loyalty economy
  • EzyCoaching → AI agent upskilling for teams and individuals
  • Evergreen → Not a website. A system that holds what you do, and keeps up as it changes.

Less AI theory. More AI in production.