About me

“If someone learns something from my mistakes, those mistakes were worth making.”

Hi, I’m John, a Canadian citizen who has been living in Germany for many years.

My career rests on a simple, well-tested hypothesis: so far, everything has worked out. Projects went live, systems kept running, and the few that didn’t became my best, if sometimes hard, lessons.

I see no reason to change the approach, so I stay optimistic and keep building.

For many years I’ve been untangling the yarn ball of project requirements and turning it into prototypes and concepts that actually run in production. I’ve done that as a solution architect, as a pre-sales consultant for clients like KfW, DB Systel, DHL, Robert Bosch GmbH, Audi, Volkswagen and T-Systems, and today as an IT system architect in enterprise networks. Along the way I learned that the most interesting and, at the same time, most challenging part of any architecture is rarely the technology. It’s the people. (The technology is a close second.)

Today my focus is research in AI and automation across on-premises, cloud and enterprise environments. That includes fine-tuning for air-gapped systems, automation pipelines, and the question of how to build AI systems people can actually trust. As an independent researcher, I test ideas where they’re allowed to fail: at home, on my own hardware, long before they reach anyone’s production environment.

Knowledge gets better when you share it. That’s why I teach, write, and publish my experiments openly on GitHub and Hugging Face. My book Application Development 2.0 came out in 2012. By now the paperback is dog-eared and quietly aging as shelfware. I, on the other hand, don’t age – quite the opposite: I’m a long-term support release. Stable, battle-tested and regularly patched.

My role model is Robert Bosch, who said he would rather lose money than trust. I think that’s the best design principle for AI there is.

Three rules I try to live by:

1. Never lose your curiosity.
2. Look beyond your own plate. (Germans say “über den Tellerrand schauen.” It’s tastier than “thinking outside the box.”)
3. If it can’t be explained simply, it isn’t finished yet.

Let’s connect: LinkedIn · GitHub · Hugging Face