Trained as a developer, 12 years in technical and strategic marketing, today a full-stack dev shipping AI agents and SaaS. Three disciplines that compound, not dilute.
AI agents, chatbots, RAG systems, MCP tools and automation. I build it for clients and my own products, and help companies get AI actually working.
Full-stack development with AI tooling. Claude Code, Codex, Cursor. I build faster and better by using AI in every step of the development process. Next.js, TypeScript, Supabase.
Tech marketing and strategy for B2B and B2C: Tech SEO, tracking, server-side tagging, channel mix, customer journey, lead management, nurturing flows. 12 years with enterprise clients in 8+ languages.
Selection from the last 24 months with clients.
RAG agent via Claude Agent SDK for a client's board of directors. Vector embeddings via pgvector, MCP connections to GA4, GSC, Google Ads and Meta. Document pipeline for status tracking and project follow-up. Self-learning system: the agent stores its own learnings back into the vector database.
GPT-Image2 and Nano Banana 2 generate professional product photos in any style, setting and context from bare product shots. Weeks of photographer coordination reduced to minutes per image.
Two tools. A prompt tracker that analyzes LLM responses for brand mentions, auto-detects competitors and runs comparisons. Works on ChatGPT, Gemini, Claude, Perplexity and more. Plus a grounded search tool that extracts the keywords LLMs actually rely on from prompt lists.
Go-to-market with AI. Validating lead lists at scale via satellite imagery: solar panels present yes/no, roof surface area, estimated installation cost. Too time-intensive to do manually, fully automated with AI.
Custom multi-agent system for a client with specific content challenges. Connected to DataForSEO MCP for keyword research, GSC for existing results, Brave MCP for web search. Produces content consistent with their branding, includes internal links and is quality checked in multiple ways.
Automatically generates presentations that look like they were made in-house. Works with pre-approved templates so every slide is consistent with the client's brand guidelines.
Shipped solo, live in production. Design to deploy.
Wedding planning SaaS. Full planning tool for couples: timeline, budget, guests, suppliers. Multilingual EN + DE.
Design · Code · GTM
Subscription management tool. Overview of subscriptions, renewals, spend tracking. Clean UI, mobile-first.
Design · Code · GTM
Sudoku game for Android. App design, development and ASO. Built out of pure interest, and because it’s fun.
Chrome extension to inspect the dataLayer during GTM and tracking implementations.
I don't collect certificates to fill a wall. Two of them say a lot about how I work today, the rest are tools I use every day.
More and more I end up working with larger companies, and they run on Microsoft. I used to build around that ecosystem and plug my solutions into it. With Azure AI Foundry I can now deploy AI the way Microsoft intended: inside the client's own environment, with the security and governance that are already in place.
Becoming a Data Protection Officer isn't the plan. But my work constantly touches (personal) data: tracking, CRM, AI agents reading customer data. I want to know exactly what is and isn't allowed, so I can think along with my clients and flag a problem before it becomes a real one.
Implementing AI is as much a change process for people as it is a technical one. I want to understand what moves people, so I can support clients on that side too.
A concrete AI implementation, a marketing strategy question, or want to spar? Send a mail, I reply within one working day.