Taking on new engagements
It usually comes from an organization that hasn't agreed on what it's building. I work on both ends of that: the flag in production, and the room where the decision gets made.
The work
Most of it is advisory, where the point is a senior answer rather than an extra pair of hands. Some of it is embedded, running engineering part-time for companies that need the seniority without the headcount.
Time-boxed work where the value is the expertise rather than the hours. Feature flagging and OpenFeature from one of the standard's maintainers, and scoping workshops for teams working out how engineering holds together once agents write the code.
ForTeams who need a senior outside answer to a specific question, and platform or DevEx groups who want depth without another full-time hire.
Head of Engineering or CTO at a fraction of a full-time hire, for companies with their first one or two engineering teams. Technical direction, hiring and growing the team, and turning delivery into something the rest of the business can plan around.
ForFounders who've outgrown running engineering themselves, and companies in the gap between two engineering leaders.
How I work
Nobody climbs harder by being careful, they climb harder because the protection is good.
Release confidence is the same trade. A team ships faster when the flag, the gate, and the rollback hold, not when it agrees to be more careful.
Principles
Each one costs me something I also want. How I got here →
The thesis
AI is stripping out the incidental friction that used to produce shared understanding as a byproduct of building software. That makes shared understanding the scarce resource, and pushes engineering work left, toward defining what a system should do before anything gets built.
This is where I work that out in public, and it is the operating model I bring to the leadership work. The series is the fastest way to see how I think, and a scoping workshop is the fastest way to try it on a real piece of work.
The model
The team decides what to build and merges that into a spec. The agent generates, reviews, and iterates at machine speed. A human accepts what reaches production.
Track record
Technical steering committee member on the CNCF's vendor-neutral standard for feature flagging, active across Java, Go, Python, and JavaScript.
Working between cloud-native projects and the teams trying to adopt them.
First cohort at the Agentic AI Foundation, the open home for the agentic AI stack.
Around eight engineers. Hiring, growth conversations, and the case for each promotion.
Architected a centralized feature flag management service across all Dynatrace clusters and stages, and wired it into OpenTelemetry.
Almost a decade in the Java toolchain: Gradle, SonarQube, CI/CD, and the plumbing that decides whether a team ships.
Sole owner and managing director of ecovas GmbH from 2012 until it merged into Adenso GmbH in 2016, building GPS tooling for waste-collection fleets, with the clients and the payroll included.
Founder and chair of a community club in Frohnleiten that runs on volunteer enthusiasm alone.
Speaking
Conferences, meetups, and internal engineering events, in English or German, delivered from inside the projects rather than from the outside looking in.
Nov 9–12, 2026
Salt Lake City
Previously ContainerDays, KubeCon EU, Devoxx, DevoxxUK, JavaCro, JCON, Cloudland, JNation, and meetups across Europe. Topics, formats, and booking →
Writing
Feature flagging, observability, engineering leadership, and where open standards are heading. Read the blog →
Contact
Send a note with what you're trying to ship, where it's stuck, and roughly when. I'll come back honestly about whether I can help, and if I'm not the right person I usually know who is.