Yu Xi Chau

Hey, I’m Yu-Xi Chau. I currently work as a Generative AI Lead at FWD Insurance, where I focus on building GenAI solutions and establishing AI Governance frameworks.

I’ve been working with data and AI for over a decade, starting as one of the early data scientists in Hong Kong. These days, I’m deep in the world of LLMs, RAG systems, and agentic workflows—building things that actually ship to production.

I have a PhD in Applied Mathematics (Complexity Science), which gave me a solid foundation for working with complex systems. Over the years, I’ve worked across insurance, real estate, retail, CRM, and gaming—the industries change, but the core challenge stays the same: how do you build AI systems that are reliable, governable, and actually deliver value?

Right now at FWD, I lead Generative AI initiatives and AI Governance work, which means building and deploying GenAI applications, setting up guardrails for responsible AI, and making sure our systems are production-ready.

On this site, I write about AI systems, MLOps, and whatever else I’m exploring. I’m also working on integrating the Zettelkasten method into my personal knowledge management workflow. Outside of tech, I’m a history nerd who reads a lot about the Ming Dynasty and Hong Kong history.

Recent Posts

The Engine Under the Hood

3 minute read

If you’ve looked at the price-performance charts for large language models recently, you might have noticed something strange. Google, the company many had w...

The Bug in the Map

3 minute read

I love it when you find a bug in the world. Not a software bug, but an error in the official story of things. A detail that doesn’t quite fit.

Which Scaffolding for Your AI Agent?

3 minute read

We’ve all been amazed by what LLMs can do. But the real work begins after the demo, when you have to make an AI do something useful. That means giving it too...

The Great Consolidation

3 minute read

I’ve noticed something when using GitHub Copilot. If I start writing some code to analyze data, it almost always suggests I use Python and the pandas library...