
Frame-Accurate Mobile EEG Stimulus Presenter
Bachelor's thesis, grade 1.0 · g.tec medical engineering
Bachelor's thesis at TU Munich with g.tec medical engineering: a React Native/Kotlin module that presents picture, text and sound stimuli for mobile EEG experiments on Android with frame-accurate timing. Onsets come from the Android FrameTimeline and AudioTrack timestamps, sit on a shared session clock and were validated against photodiode and AUX hardware references. Graded 1.0 (best possible grade).
- Research
- Android
- Kotlin
- React Native
- EEG
Live interview trainer: practice tech interviews in real-time voice conversations with an AI interviewer, generate sessions from job descriptions, solve coding, SQL and whiteboard exercises in the browser and review structured, scored feedback. Built with Next.js 16, Convex and the Gemini Live speech-to-speech API.
All-in-one event platform: visitor mobile app, organizer and admin dashboards, Stripe Connect ticketing with signed-QR admission, photo memories, subscriptions, friend location sharing and Eventure Go, an event-based ride-hailing product with driver app and fleet dashboard. Modular monolith on Convex with end-to-end type safety across Expo and Next.js.
- Mobile App
- Web App
- Convex
- Stripe
- React Native

Ecommerce built with NextJs, Stripe and Sanity

News Application built with NextJs 13 and TailwindCSS

Cloudflare Serverless Worker
This Hono application is a serverless REST API designed for Cloudflare Workers, utilizing Deno's Hono framework.

Car Reinforcement-Learning
Reinforcement-Learning program from scratch without libraries

Computer learns Snake with an Reinforcement-Learning-Algorithm