At ten I took my father's computer apart. I wanted to see the thing inside the box, the part that made it work. I got it back together, mostly. The curiosity stayed.
That curiosity is the thread through everything since: writing software, drawing system boundaries, leading teams, teaching kids and students, giving talks, building companies, and now working out what AI means for the craft.
Where it started
I studied Business Informatics at NORDAKADEMIE as a dual student, which meant university and OTTO at the same time from 2009. OTTO rotated me through a lot of the company: controlling, SAP HR development, e-commerce, corporate strategy. I also helped get the startup incubator Liquid Labs off the ground, where Risk.Ident later spun off into its own company.
I learned something there that shaped everything after. The interesting problems are rarely only technical.
OTTO: learning to ship
In 2013 OTTO took me on as a software engineer. I joined a SCRUM team of ten building a new backoffice system for otto.de.
Then a colleague and I did something I am still proud of. We convinced leadership to start a new team focused on social media features, and I co-led it technically. Two weeks in, a small feature shipped from that team generated revenue in the three-digit millions.
That taught me how much a tiny, well-placed piece of software can move, and how much of that depends on a team being allowed to build it.
Breuninger: architecture as a way to keep moving
From 2017 I was Lead Software Architect at Breuninger, responsible for the online-shop platform. Thirteen teams worked on it, 40 internal and 30 external developers.
In those years we relaunched the shop in Germany and Austria, expanded into Switzerland, and added a marketplace. Online order revenue grew 35% in 2018. We did it without piling up the kind of technical debt that quietly slows everything down later.
The lever was architecture. I established a vertical architecture that made teams customer-focused, data-driven, independent, and able to organize themselves. Breuninger is where I learned that good architecture is mostly about making change cheap.
BLUME 2000: the whole picture
As CIO at BLUME 2000 from 2020, I owned the e-commerce IT department. The team was not set up for what was coming, so we rebuilt the product, organization, and technology layers, and shipped a new online shop on a new foundation.
CIO work taught me where technology, organization, operations, and business responsibility actually meet. In small daily decisions, not in strategy slides.
TalentFormation and hackers&wizards
After BLUME 2000 I moved into company rebuilding at TalentFormation, working on architecture, delivery, and verticalization. That is where the first real shape of my agentic engineering work appeared.
Today I run hackers&wizards, my company for agentic engineering: training, workshops, and practical AI-assisted software development. It has its own home at hackersandwizards.dev.
The teaching thread
Teaching keeps showing up. Hacker School since 2014, university lectures, mentoring, conference talks, workshops, and now working with AI systems.
The pattern is the same every time. If you want a person or a system to do good work, you have to make context visible. Examples, feedback, boundaries, patience, and enough structure that the next step becomes obvious.
That is why mentoring and agentic engineering feel like the same work to me. Getting good results out of AI is a context problem and a teaching problem before it is a tooling problem.
The question I am working on now
How do we use AI to make software teams more capable without losing craft, judgment, or ownership?
That question drives hackers&wizards, my talks, my writing, and most of the conversations I have with CTOs, developers, and educators.
If AI makes producing code cheaper, teams need stronger architecture and clearer thinking, not less of either.
The same curiosity that made me open that computer still runs the work: understand the system, make it useful, teach what you learned.