Hello, I'm Sanhe

Hello, I'm Sanhe

If you're interested in AI, software architecture, and building in public — subscribe to my blog and join me on this journey.

This is my first post on ghost.io platform. Instead of writing a polished "About Me" page, I want to have an honest conversation about who I am and why I've decided to start writing publicly.

A Career That Started From Zero

When I graduated, my coding ability was zero. Literally zero.

Then I stumbled onto a path I never planned: Data Scientist → Data Engineer → Data Architect → Enterprise Solution Architect → AI Architect. Every transition was forced by circumstances — I was pushed forward more than I chose to move. But in the end, every step was still my choice.

During my time at Amazon, Customer Obsession became deeply embedded in how I think. I discovered that my greatest satisfaction doesn't come from writing elegant code — it comes from helping customers solve real problems and realize their value.

Later, I built two small businesses on my own, both worked out. Without AI, there's no way I could have handled legal, finance, accounting, compliance, tax, customer support, customer engagement, sales, and marketing all by myself. Ironically, for a tech person like me, building the product was the easiest part.

The Things I Did Before the Industry

This is a bit awkward to say, but I did build some things before they became mainstream:

  • 2020: Built a ChatGPT-like Q&A chatbot using Amazon's ecosystem — before ChatGPT existed. I needed a way to search and answer questions from my massive collection of blog posts and technical docs
  • Early 2023: Got fed up with manually syncing codebases and copy-pasting context to AI. Built my own context sync engine that kept AI updated with real-time codebase state — essentially what Cursor does. Finished it about 2 months before Cursor launched
  • 2024 when MCP launched: Built Google Gmail/Drive MCP servers about six months before Google's official release
  • March 2025: Realized MCP loads too much context, built a multi-layer lazy-load solution for prompts and tools — six months later, Spro turned a similar concept into the Agent standard
  • May 2025: Built a mini kanban system to sync tasks across multi-agent workflows — six months later, Anthropic shipped Claude Code's Task system with similar functionality

You might think I'm bragging. But that's not why I'm telling you this — quite the opposite.

Am I a Loser?

I could see these things coming. I actually built them. But I never turned any of them into something big.

Never made them into products. Never built companies around them. Never scaled them to impact more people.

I had the vision. I took action. But I didn't follow through. Honestly, this is something I regret.

I have 190+ open source Python packages with roughly 10 million monthly downloads combined. I have 700+ GitHub repos — mostly POCs and sanitized versions of enterprise solutions I've built. I've slowly written over 2,000 blog posts over the years.

But all of this was done alone. I was absorbed in my own world. Never built a community. All my businesses were solo operations. I was always building privately.

Build in Public

So I've decided to change.

I know I can't become a YouTuber right now — that's too hard for me at this stage, and I have a lot of work on my plate.

But I can start with writing.

This blog is my first official step toward building in public. I want to share things I believe are valuable — technical insights, entrepreneurial lessons, and hard-earned wisdom from years of grinding in this industry.

I won't pretend to be an expert. I'm just someone who keeps tinkering, keeps learning, occasionally sees things before others do — but fails to seize the moment.

If you're walking this path too, let's walk together.


Subscribe at blog.sanhe.me — Let's build in public.