programming as the greatest hobby

➜ contact

My GitHub and Telegram accounts.

➜ projects

in development

A new portfolio with an improved design is in progress.

Wildlife Camera

Made with a Raspberry Pi, a night-vision camera and a motion sensor. Captured images are sent to my server over a WLAN network (using Python) and then displayed on a NextJS website.


A vocabulary trainer made with NextJS, Prisma ORM, MySQL, NGINX, and Docker. I've been working concentrated on it since December 2022.

in development

Nextcloud is slow and not stable. That's why I need an alternative. I am occasionally developing Lastcloud with my preferred stack. The hardest challenge was to implement file uploads, for which I am using octet streams and no forms.
I have chosen this name because it's the last cloud anyone is ever going to need, though I came up with the idea while I was texting with someone, and, of course, the name is more of a joke.
What makes Lastcloud special is that I am trying to make full use of my ZFS RAID by exposing options such as deduplication and synchronous writes to the user.
With Lastcloud, I have learned a few things already: Tailwind CSS, NextJS App router, better Database organization, and last but not least the lightning-fast Sharp library that I use to compress images.


A real-time chat application similar to Discord. I am developing it with a friend, and much lies ahead yet. Frankly, not much has happened the last weeks, and I am exhausted. It was a little too complex to do with just two developers, and numerous bugs arised. Perhaps we are going to start over, or continue when we've got the strength again.


The source code is in my GitHub repository.

AI Dog Detection

Searches for our house's dog in images. Might eventually build a website 'findMyDog' where my model is run on images uploaded by the client.

MCAltChecker (stale)

AI written in Tensorflow (Python) to check if a Minecraft Player is cheating or not using server's APIs to inspect their game stats. The frontend developer lost interest (for the moment), however the AI guessed 75% correct of a small dataset of 100 players.

CLI pythonVocabularyTrainer (old)

My first attempt of writing a vocabulary trainer. Only worked on the Linux Command-Line-Interface, but with colors to improve the experience.


administration of a public self-hosted Nextcloud, data stored on ZFS (Zettabyte-File-System)