I am a business management graduate (B.A.) who has discovered a passion for coding while digitizing a family business. What started as a simple online course during a summer vacation turned into managing a full on-premises environment and developing full-stack applications.

I have spent the last few years engineering solutions for our family butcher shop (~20 employees). I've built everything from AI-orchestrated support bots to automated digital signage and paperless office systems.

So, why look for something new? I love building complex automations. However, in a small enterprise, these powerful tools are triggered maybe 3 times a day and I just feel like the projects I thrive on are bigger than that. I'm at a point where I've automated about 80% of my original office work, and I feel the effort of building additional tools doesn't create that much more impact anymore. 👉 I am looking for meaning and a new challenge. I want to build systems that run thousands of times a day. I want to apply my skills in an environment where efficiency, iteration, and well-crafted code create measurable impact and delight for users.

My toolbox combines modern full-stack development with deep system administration and DevOps knowledge.

"AI is my co-pilot, not my autopilot."

My daily driver is VS Code, though I frequently utilize JetBrains IDEs when I need deep introspection. I love using AI tools—GitHub Copilot, Gemini CLI, and AI-Enhanced VS Code Forks—to accelerate parts of development and enhance learning.

• Docs shortcut: I use AI to skip long documentation lookups and speed up routine tasks.
• Logic is human: For safety and security, I design the logic and control flow myself. I use AI to implement the specific lines, but the "blueprint" remains mine.
• The "Gutscheineland" Lesson: This was my first project driven heavily by AI Agents.
  • The Win: Incredible speed to MVP. (The strongly typed, configuration-heavy nature of Payload + Next.js is great for AI-assisted workflows.)
  • The Trade-off: AI tends to generate a "salad bowl" of code structures.
  • The Job: I shifted from writing to reviewing and refining. I focus on keeping project structure maintainable and the product delightful.

Building a SaaS-style voucher platform.

• Tech: Next.js, PayloadCMS, PostgreSQL, Stripe Payments.
• Ops: Deployed via Coolify on Hetzner, using S3 for asset storage.
• Workflow: Full CI/CD implementation via GitHub Actions.

I architected a complete on-premises solution to modernize our business operations.

• Infrastructure: Run on a powerful Mini-PC via Proxmox (LXC Containers + VMs) + NAS.
• Networking Security: Tailscale serves as the backbone—used not just as a VPN, but as a DNS orchestrator (with Pi-hole sidecar) to ship tools internally via FQDNs. I use Tailscale Funnels to securely share services (like Nextcloud) without dangerous port forwarding and absolutely love the MagicDNS combined with the "tailscale serve" reverse proxy and the brand new "advertise service" features.
• Backup Strategy: Automated backups via Proxmox Backup Server to NAS ➔ Snapshotted via Kopia to a remote location (3-2-1 strategy).
• The Paperless Office: Deployed Paperless-ngx managing 15,000+ documents.
  • Custom Python Scripts: Developed tools to scan directories for missing invoice numbers, generate ASN barcodes, and export CSVs with formatted POS data specifically for DATEV accounting import. (View Scripts)
• Open Source Contribution: I contributed the emergency recovery mode to Kopia (View PR) to allow recovery when repository storage is fully exhausted. (WIP)

• Stack: Self-hosted Chatwoot (Hetzner), Nginx Reverse Proxy, Postfix (Mail Relaying).
• Automation: Integrated n8n via webhooks/API to manage WhatsApp newsletters.
  • Logic: Handles subscribes/unsubscribes via contact tags and sends automated welcome coupons with comprehensive logging for compliance.
  • Scaling: Implemented Redis for rate limiting to ensure stability.
  • AI: Currently orchestrating an AI Agent to handle Level 1 customer inquiries. (WIP)

• The Problem: Manually updating menu screens was slow and error-prone (took me approx. 2h a week).
• The Solution: A Python pipeline that extracts data from the central planning Excel to JSON, renders it into an HTML template, zips the package, and pushes it via the PiSignage API to POS displays.
• Result: Zero manual intervention required for daily updates.

"I can balance a binary tree, but I prefer building real-world software."

• Lecturer: DHBW Heilbronn (Food Production & Sales) | Member of the Quality Circle for Food Management.
• Academic: B.A. BWL Food Management (ECTS classification: A).
• Computer Science:
  • CS50x & CS50w: Completed Harvard's introduction to CS and Web Programming.
  • FernUniversität in Hagen: Guest auditor for Go (Imperative Programming) and Java (Algorithms & Data Structures) to understand the theoretical underpinnings (Binary Trees, Big O, etc.).
• Teaching: Private tutor for Maths, Physics, Chemistry & Economics.

"Impossible isn't in my vocabulary."

The only variables are effort and cost. I believe the best solutions emerge when traditional values meet an innovative mindset. I've learned that with a computer, I can build anything—and that's what really gets me addicted to coding.

Let's connect and create something new! 💡