Head of Software Development | AI Enthusiast | Tech Author
A department head who still codes
My journey in software development began in 2009 at COSYS Ident with mobile data capture devices and C# .NET Compact Framework. Today, I lead software development at WERBAS KSR with a team of 20 developers.
Between these two points: positions at Siemens, Trafag and Werbas. Always deeper into architecture, always broader in responsibility. From developer to team lead to department head - but never away from the code.
When I'm not developing or leading, I fly FPV drones through factories and over racetracks. Or I play with my dog Luna.
Three Pillars of My Work
Leadership & Transformation
Servant leadership, Team Topologies, and the courage to transform.
Leading 20 developers through a major transformation -- not by directing, but by serving. Migrating a C# WinForms monolith to a modular SaaS platform while restructuring into stream-aligned teams, a platform team, enabling teams, and a complicated-subsystem team. My leadership philosophy: remove obstacles, foster ownership, invest in people's growth, and trust teams to solve the right problems.
AI & Innovation
Developing AI services, establishing LLM workflows, building the future.
Leading an R&D group focused on AI integration. Built a central AI service running on Azure that handles text processing, speech-to-text, document extraction, and translation in 17 languages - consumed by multiple applications via API. Also: Lecturer on 'AI and EU AI Act', member of LLM research group at University of Cologne.
Knowledge Sharing
Writing, teaching, researching.
Contributing to the developer community through technical articles in dotnetpro and windows.developer magazines. Lecturing on AI and EU AI Act. Participating in academic research on LLM cooperative behavior at the University of Cologne. Serving as IHK examiner since 2012.
My Journey
Head of Software Development
WERBAS KSR GmbH
20 developers, 1 vision: Legacy to Cloud
Head of Software & IT
Magnetic Sense GmbH (Trafag AG Group)
Sensor technology meets software leadership
Senior Software Engineer / Trainer
Trafag GmbH
R&D focus, first leadership responsibilities
Software Engineer / Trainer
Siemens Industry Software GmbH
Enterprise-scale product development
Technologies
I don't just write code
Technical articles in leading German developer magazines.
From Blueprint to Workpiece (Based on the GEEF Principle)
November 2026A technical article on dynamically assembled multi-agent pipelines, self-improving system operation, and an open-source reference implementation.
From the advisor in the background (+ GEEF Principle)
October 2026Anthropic’s Advisor Tool adds a strategic advisory role to AI pipelines alongside execution and review. Within the GEEF principle, this role mainly helps identify grounding errors early — making the right distribution of roles and models the decisive factor.
From Prompt to Pipeline: The GEEF Principle
August 2026Single-agent AI handling generation, review, and formatting reduces quality. GEEF solves this via role separation: specialized agents iterate through Grounding, Execution, Evaluation, and Finalize, with a feedback loop until no critical findings remain. It applies from software to compliance, with open challenges in convergence, cost, and reviewer calibration.
Default Interface Methods in C# 8.0
October 2025A language extension under scrutiny. Exploring the practical implications of default interface implementations in modern C#.
Read Article →Education & Certifications
Academic
Professional
Certifications
Education
Code That's Free
Some of my work is available for everyone.
Geef.Atelier is a platform for structured text creation, combining briefing, drafting, and iterative review in a transparent workflow. It helps users create consistent, audience-tailored texts with less effort while keeping full control over tone, content, and quality.
GEEF is an AI workflow framework with four phases: Grounding, Execution, Evaluation, and Finalize. It repeats execution until reviewers approve the result, with policies for iteration limits, timeouts, stagnation, and critical errors.
A lightweight C# library that simplifies dependency injection using default interface implementations. Enables container-agnostic service registration and resolution.
A C# tool using the OpenAI API for interpreting complex data like error stack traces. Simplifies troubleshooting and error resolution by processing and clearly presenting results.
A C# library that allows you to manage, register, and execute methods dynamically. Provides functionality to find and execute the most suitable method based on provided parameters.
The Person Behind the Profile
Exploring nature with Luna
My Labrador Luna keeps me grounded. The best ideas come between fetch sessions and muddy paws.
View Gallery →
Lecturer for AI Competence
Since February 2025, according to Article 4 of the EU AI Act, companies must demonstrably ensure that their employees possess AI skills. My employee training: 90 minutes of training, online exam, and certificate.
Visit →
FPV Flying
Racing through the sky at 150km/h. Sometimes work, always fun.
Visit →
Life in Pixels
Occasional snapshots from life beyond the screen.
Visit →Let's Talk
Whether it's a project, a talk, or just tech chat - I'm always up for a conversation.