I've been taking computers apart (and occasionally putting them back together) since I was 12, when I was compiling FreeBSD kernels and reverse-engineering TCP/IP stacks for fun. That obsession with understanding how things actually work never went away - it just evolved into building cloud infrastructure, automation pipelines, and AI systems.
These days I work at the intersection of DevOps, Cloud, and AI. I help businesses automate their workflows, and I build my own production systems where uptime and cost actually matter.
k0sdev-lab - Provisions a Hetzner 2 node k0s cluster labratory via Terraform.
Sovereign-Mesh - A secure, multi-tenant AI PaaS on Hetzner Cloud. Orchestrates private departmental RAG enclaves using vLLM, Qdrant, and TEI on a Zero-Trust K3s cluster. Features CKS-grade isolation (Restricted PSA, NetworkPolicies) and automated MLOps GitOps workflows. Built with Terraform, Kubernetes, ArgoCD, vLLM, and LangChain.
Tokenguard Operator - A Kubernetes operator that scores ServiceAccount least-privilege in real time by comparing granted RBAC permissions against actual audit log usage. Detects external IP token abuse for supply chain attack mitigation. Built with Go, Kubebuilder, controller-runtime, and GitHub Actions.
NinjaDevOps - An interactive DevOps challenge platform with 23 real-world Linux emergencies (Docker, Kubernetes, DNS, etc) in live Google Cloud VMs. Debug broken servers, fix misconfigurations, and deploy under pressure. Built with React, Firebase, Google Cloud Run, Compute Engine, GitHub Actions.
Hetzner-Kubelab-Pipeline - Deploy 1 master node and 4 worker nodes on Hetzner, with a basic application + CI/CD pipeline with ArgoCD + Prometheus + Grafana
Terraform-K8s-Kusama-Validator-Project - A cluster of Kusama validators that can expand / contract
Video / Song Generating SaaS - A serverless AI video generator SaaS with Frontend. Send a text or voice message, get a video back. Built with Google Cloud Run Functions.
Content Machine - A serverless AI video generator. Send a text or voice message, get a video back. Built with AWS Lambda and a multi-agent AI pipeline.
ChatOps Static Site Platform - Deploy websites by texting a Telegram bot. Built in one day using patterns I'd refined over months. Costs essentially nothing to run.
Serverless Transcription Bot - Voice notes to text via AWS Transcribe. 100% serverless, zero idle costs.
I also maintain production RAG bots and n8n workflows for a German eCommerce agency - real systems, real users, real consequences when things break.
- Certifications: AWS Cloud Practitioner ✓, AWS Solutions Architect Associate ✓, Certified Kubernetes Administrator - CKA (March 2026), Certified Kubernetes Application Developer - CKAD (March 2026), Certified Kubernetes Security Specialist (CKS), KCNA + KCSA - Kubestronaut.
- Stack: Terraform, Kubernetes, AWS (Lambda, S3, CloudFront, Bedrock), Python, n8n, Docker, GitHub Actions
- Philosophy: Infrastructure as Code, GitOps, cost optimization, and building systems that serve humans (not the other way around)
If you want to dig deeper, here are some older experiments:
- Instagram Lead Hunter AI - AI-enhanced lead qualification
- YouTube → Medium Agent - Automated video-to-article pipeline
- YouTube Comment Classifier - MLOps pipeline with FastAPI + K3s
- TCP Hijacker Classic - Educational networking/security tool (a callback to my early days)
Currently: Working with my freelance clients, preparing for CKA, running KubeLab, and exploring for DevOps/Cloud Engineering opportunities in Germany.