Profile

Grade: 6.0

For my Matura thesis, I researched and implemented linear and differential cryptanalysis. The goal was to build a deep understanding of the mathematical structures behind modern encryption methods. I wrote a Python program that models a custom Substitution-Permutation Network (SPN) and simulates attacks using both analytical methods to compare their data economy and success rates.

Check out the Source Code on GitHub

My proudest creation yet. Its an iOS native app that helps people cut down screentime using a simple mechanic: roll a dice each day to earn time on distracting apps. Built on Apple‘s Screen Time API. Currently live in the AppStore with 500+ active users.

Check out the Website

Phantom is a new kind of AI interface for the Mac. Instead of living in a chat window, it sits in the notch, understands what’s on screen, and operates apps by clicking, typing, and navigating like a person. It can execute multi-step workflows autonomously in the background or collaborate with me in real time.

I built Phantom to explore what comes after chat: AI as an interface layer for the entire operating system. It’s a tool I use daily.

Check out the Website

My first app ever. It all started with the desire to build something of my own and, in the process, learn a new skill. I chose a simple format to train myself, and thus Cali was born: a simple, intuitive calorie tracker that uses AI to detect the nutrients in your food. It also features an integrated AI chat where you can ask a personalized nutritionist questions or even have a custom diet plan created for you.

Check out the Website

I traded ICT for five months, achieving a live trading ROI of 21.7% during the first three months trading Solana on Bybit. Alongside live execution, I extensively backtested multiple strategies on demo accounts, taking a data-driven approach to refine and iterate my methods.

I explored algorithmic trading by developing simple strategies, running Monte Carlo simulations, and creating a machine-learning-based indicator algorithm designed to self-optimize. Ultimately, I chose not to deploy these systems in live market environments because the simulations did not properly account for real-world factors like slippage and trading fees.