Senior engineering meets custom skills, automation and multi-agent orchestration. „AI writes code" becomes production-ready software — traceable, tested, maintainable.
What AI does to our work, our thinking and our professions — experience and essays from practice, no hype.
Current experiences from working with Claude Code — short notes, decisions, aha moments.
To the blog → // EssaysLonger pieces on the big questions around AI in software development — more in depth than the blog.
To the essays →Prefer email? Subscribe to the newsletter →
I combine solid software engineering with a system of prompts, skills and orchestration — so that AI output becomes reliable, production-ready software.
I have been building software for over 30 years. Since the earliest days of Claude Code I work with it daily — on refactorings of large codebases, new features, test suites and my own developer tooling.
My approach is not „prompt more", but better systems. More about me →
Depending on the task I work manually, semi-automated or fully orchestrated. The art is knowing which level when.
Closely guided pair engineering — I think, plan and review at every step. Maximum control for delicate changes.
Clearly scoped tasks run in controlled loops: plan, build, test, review — I keep my hand on the wheel.
Several specialised agents work coordinated and autonomously on large efforts — with budgets, guardrails and verification.
I get your team up to speed — hands-on, on your real tasks.
I set up Claude Code for your codebase and processes.
We build together — at the pace of an AI-assisted team.
I build your product — largely AI-assisted. RAMP is the proof.
My experiences with Claude Code form a web: one common thread, many cross-connections.
Whether it's setup, concrete delivery or a workshop for your team — drop me a line about what you have in mind.
