WenHao Yu余文皓 · Based in Taiwan

I spent twenty years in tech, at Nokia, Novatek and AUO: seven as an engineer, ten as a product manager, and three in data analysis as staff to a business unit head. As a product manager I ran phone projects on tight schedules and automotive electronics on long, strict ones. Product planning, business decisions, getting departments to agree: I have done each of them.

The longer I did it, the more I felt I was standing still. My skills kept growing, but most of each day went into getting other people to agree to let me do the work.

In 2019 the company moved its whole phone division to China. I chose not to go, and ten years of contacts and standing went almost to zero. The next year I started keeping a diary. The question I wrote on the first day was: “I know the life I have now is not the one I want. How do I find the courage to leave my comfort zone?”

In 2023 I went back to school.

I studied decision analysis at Minerva University. In 2025 I left my company, the year Claude Code came out. I was among its first users and started building products with it: AI Resume Advisor in four months, NeatToolkit in three days.

At first I only used it to build products. Then my goals, my daily work, what I learn, my clients and my life at home moved onto it, one piece at a time.

The way I work with AI comes from those twenty years.

A project used to start with a Project Charter, a document that says why we are doing this and where it ends. I still write that first, for a product or for a quarter's goals.

Defining the problem, breaking down the requirements and accepting the result is what a product manager does every day. It used to be with engineers. Now it is with AI.

As staff to a business unit head, my job was to find what the data was saying and turn it into a strategy recommendation. I run my own business the same way.

Inside companies I led the rollout of Jira and Polarion and set the processes for project management, problem solving and change control, so I know where the plan and the daily reality come apart. I use the same experience when I help a team bring AI into its process.

I wrote all of it down.

Readers asked less about the products and more about how I work with AI: what my setup looks like and how I arrange the work. So I kept writing about that.

Then people wanted to learn it and companies wanted to adopt it. I started teaching, went into companies to work with their teams, and founded Confluence Partners with a few partners. In August 2026 I passed Anthropic's Claude Certified Architect exam. All of this became my income, without my planning it that way.

Doing work I like, with my time my own, and being useful to other people: this is the life I wanted. The question on the first page of my 2020 diary has its answer now.