I built a stock screener which uses my other stock analyzing tool built with Next.js. This app is a classic SaaS.

Homepage: Stock Screener.

I built this site using my own lightweight framework called StackZero, which is based on a simple and efficient stack: Node.js with Fastify, htmx and SQLite — all wrapped in Docker and deployed via GitHub Actions. The entire setup is designed to be low-cost (under 5€ per month, including DNS via Ionos and hosting on a Hetzner server), transparent, and fully under my control.

GitHub: StackZero.

I setup an home server using an HP ProDesk 600 G4 running Ubuntu Server, fully automated with Ansible and containerized via Docker Compose. It hosts a media automation suite based on the arr stack and Plex — all using images from LinuxServer.io. The setup is zero-config, reproducible, and entirely under my control — ideal for managing media and services on my own terms.

I migrated our infrastructure from AWS to Hetzner, reducing monthly costs from ~4.000€ to ~100€ — saving around 3.900€ per month. The setup previously relied mostly on EC2 and ECS, which I fully replaced by consolidating all services into a single Docker Compose stack. I added NGINX as a reverse proxy and automated HTTPS via acme. Everything is now containerized, reproducible, and fully under my control — deployed on a performant Hetzner server.

I built a stock automation tool in Python based on the PowerX Strategy by Markus Heitkoetter. Designed as a command line application, it generated buy and sell signals to support my trading decisions — all fully local and under my control.

GitHub: PowerXStocksAnalyzer.

I built a small forecasting tool in Python to predict future weather based on historical data. It allows data updates via the WorldWeatherOnline API and supports feature selection and training via a simple CLI — ideal for experimenting with time series prediction in a fully local setup.

GitHub: Weather-Forecast.