Backend-leaning software engineer with 6+ years across research, industry, and independent practice — from Fraunhofer media/streaming pipelines to fintech trading platforms. I build production backends, data pipelines, and self-hosted platform infrastructure that scale, and I use AI/ML to automate the workflows around them.
Not scripts — systems: reliable, observable, and maintainable. M.Sc. Computer Science, TU Berlin · based in Berlin, Germany 🇩🇪
backend: [Python, FastAPI, Go, Node/TypeScript, REST, WebSocket]
data: [PostgreSQL, MySQL, Redis, MongoDB, ETL & streaming pipelines]
platform: [Docker, Docker Swarm, Kubernetes, Ansible (IaC), Traefik, CI/CD]
observability: [Prometheus, Grafana, centralized logging, automated backups]
cloud: [GCP, AWS, Azure, GitHub Actions, GitLab CI]
ai_automation: [LLM routing, RAG/agents, MCP tool integration, workflow automation]| Project | What it is | |
|---|---|---|
| mlflow-selfhosted | Self-hosted MLflow tracking server (MinIO + Postgres) — reproducible MLOps infra |   |
| homelab-compose-stack | Reusable Docker Compose templates for self-hosted homelab infrastructure |   |
| cftc-cot | Python library to download & parse CFTC Commitments-of-Traders data |   |
| cot-analysis | COT feature engineering & market-sentiment / directional-bias scoring |   |
| JobScraper | FastAPI microservice scraping job listings → MySQL → REST API |   |
| termin-bot | Berlin Ausländerbehörde appointment-finder bot |   |
| Image | Pulls |
|---|---|
comfy-ui |
23,039 |
jobscraper-api |
1,758 |
comfy-download |
717 |
jellyfin-media-status-sync |
275 |
jupyter-gpu |
243 |
mlflow |
112 |
ffmpeg |
0 |
Updated daily from Docker Hub's public API.
- Platform & MLOps — self-hosted stacks (Docker Swarm, Ansible), MLflow, monitoring
- Automation-first infrastructure — reproducible, observable, self-healing systems
- LLM routing & multi-model orchestration for engineering workflows
⚡ Build systems that run themselves.