Hey, I'm Willem Meints

I've been working as a software engineer at Info Support in Veenendaal since 2007, and later switched to my role as software architect. In the earlier years, I did what architects and software engineers are supposed to do: design systems, guide teams, and give the occasional presentation. Later, I discovered AI and learned to build neural networks and use generative AI. I wrote books on both topics, the most recent being “Building Effective LLM-based applications with Semantic Kernel.”

A lot changed for me when I first saw GitHub Copilot in action during an MVP Summit. Not because it was spectacularly good—it made plenty of dumb mistakes — but because I saw potential. For the first time in years, it felt like software development was about to undergo a fundamental change.

What drives me

I'm tired of the AI hype. Consultants promising that AI solves everything and will replace developers. It’s not going to happen for many more reasons than I care to cover. Of course, the demo works. But it’s a bit like McDonalds. You’ll get hungry after 20 minutes.

That's why I focus on what actually works. I help a team at PostNL build multi-agent systems on AWS with Claude. I help Rabobank adopt Open-Source. I run workshops on coding agents with Semantic Kernel for C# developers who don't want to mess with Python notebooks. I interview people about how they actually use AI assistants — not how they think they should use them. And you can find me at the Info Support HQ office in Veenendaal, helping many more teams run their AI projects in my role as Chief AI Architect.

I don't believe in pretty PowerPoints and shiny YouTube demos. I believe in working code and people getting better at their craft.

What I do

My work revolves around one question: how can developers use AI without it feeling like yet another tool that creates more problems than it solves? That means context engineering—teaching teams how to make their codebase comprehensible to AI. Multi-agent systems—getting multiple specialized agents to collaborate instead of one "do-everything" chatbot. And mostly: experimenting, failing, learning, and sharing what works.

I work with Python, C#, and TypeScript. Usually backwards: figure out the goal first, then make a plan. Most "AI strategy trajectories" start with the technology. I start with: what are you actually trying to solve?

Outside work

I live in Veenendaal with my family. I have a 3D printer that I use to make things nobody needs but look cool. I enjoy cooking, especially with my BBQ. And I spend too much time tweaking development tools—Neovim configs, terminal setups, that sort of thing. I ride a wheelchair because my body just doesn’t work right anymore. But it is the best tool I have ever put my hands on.

If you want to email me about an "exciting freelance opportunity", don't. If you want to talk about how your team can effectively use AI tooling, or just discuss multi-agent architectures or agentic AI, I'd love to hear from you.