Projects

AboutProjectsPosts

Here are some of the projects I’ve undertaken, many of which were created for personal enjoyment or as part of my studies at Hochschule Luzern. Most of them you can find on my GitHub @severinlindenmann

Magyul – Hungarian Language Learning App

React · TypeScript · Firebase · PWA

A multilingual vocabulary trainer with conjugation drills, offline-first study modes, and bite-size grammar guides.

  • Flashcards with HU / EN / DE context sentences.
  • Installable PWA experience with IndexedDB sync.

Swiss AI Weeks Banking Client Simulation

Python · Data Simulation

Synthetic banking datasets and personas that stress-test AI systems for fraud detection and client analytics.

  • Scenario engine for configurable customer behaviour.
  • Probabilistic event generation for analytics sandboxes.
Road traffic segmentation

Road Traffic Segmentation

Computer Vision · PyTorch

Semantic segmentation experiments that classify roads, cars, and pedestrians for smarter mobility analytics.

  • Training notebooks with augmentation and evaluation tooling.
  • Deployable inference scripts for batch and live use.
Wuchemenu.ch

Wuchemenu.ch

React · Contentful

Weekly recipe inspiration with curated menus, seasonal flavours, and grocery planning prompts.

  • Card-based browsing with responsive layouts.
  • Headless CMS workflow for publishing fresh menus.
One Class SVM dashboard

One Class SVM for Spam Detection

Streamlit · Machine Learning

An interactive Streamlit dashboard that invites teams to explore one-class SVM models for spotting anomalies.

  • Configurable training runs with live metric visualisation.
  • Dataset upload helpers for rapid experimentation.
Traffic accidents in Switzerland

Traffic Accidents in Switzerland

Data Visualisation · GIS

Interactive dashboards that surface accident trends across time, severity, and location between 2011 and 2021.

  • Exploratory filters with responsive map visuals.
  • Storytelling layouts for workshops and policy reviews.
Twitter classification bot

Twitter Classification Bot

NLP · Automation

A live sentiment bot that labels tweets by mood to surface community signals in real time.

  • Two-stage transformer inference with queue orchestration.
  • Dashboards to monitor sentiment trends and model drift.