How I think about AI
AI is a tool, not a goal. Every project starts with a business need, then the simplest approach that delivers. Sometimes that approach is no AI at all, and I will say so.
My AI work
Alongside business software, my focus is now AI integration. These examples come from my own work:
| Example | What it means for you |
|---|---|
| Private RAG prototype | Answers from internal documents, with data kept on local hardware |
| Provider-independent RAG application | Local or cloud AI with one setting, and "I don't know" when no source fits |
| AI test workflow | Software tests created 2 to 3 times faster in my daily development work |
| Custom AI skill | Repetitive translation tasks in web applications handled automatically |
My approach
Every solution is designed with costs and data security in mind. Careful planning keeps token costs low and sensitive data under your control.
- Integration, not rebuilding. AI joins the systems you already run.
- Private when needed. Local AI models keep sensitive data in-house.
- People stay in control. Your team approves the steps that matter.
Credentials
My AI skills are backed by certificates from IBM (RAG and agentic AI), Vanderbilt University (advanced prompt engineering) and the University of Pennsylvania (AI for business). They build on more than a decade of software engineering for German companies and six years of running my own freelance business.
Let's start small
Whether you have a clear project or only a first idea, I am happy to talk it through.