Portfolio

Welcome to my portfolio! Here, you'll find a selection of my most significant projects and achievements. Each project reflects my passion for computer science and my commitment to creating innovative and efficient solutions. Feel free to explore the different works presented here to learn more about my skills and experience.

FinalWar

Browser-based MMORPG (2006-2008) Stack: PHP (Legacy), HTML/CSS, MySQL.

The Pitch: At 15, frustrated with school, I wanted to create my own universe. FinalWar was an online management and strategy game based on the Naruto manga universe.

The Technical Challenge: It was the era of 'everything by hand'. No modern frameworks, no AWS cloud. I had to learn to manage:

A MySQL database that had to handle simultaneous actions from multiple players.
Inadequate hosting (initially a non-web server) that had to be optimized.
Maintaining player engagement over time.

The Lesson (Post-Mortem): The project was a success in terms of reputation (and traffic for my age), but it hit a legal wall: intellectual property. Not having the rights to the Naruto license, and facing the potential threat of royalties, I had to shut down the service. This project is the 'archaeology' of my journey: the code is lost, but it taught me that being a developer also means understanding digital law.

OniMemory

Mobile puzzle game & UX experimentation Stack: Flutter, Dart, Krita (Graphics).

The Pitch: Initially designed to showcase my brother's drawings (based on Asian 'Oni'), this project aimed to create a dynamic memory game on mobile.

The Implementation: I chose Flutter for its cross-platform capability. The challenge wasn't so much technical (memory game logic is simple) but related to user experience (UX): how to make the game fluid, beautiful, and 'satisfying' to play? I worked on animations, transitions, and 'game feel'.

The Ethical Pivot: During development (10 months), I successfully integrated very effective retention mechanics (reward loops, calculated frustration) to maximize future monetization. This is called Dark Patterns. When I realized I was coding a tool to manipulate players' dopamine instead of offering a healthy experience, I reversed course. I removed all toxic mechanics to keep only the pure game. Status: Under redesign for 'clean' publication (without predatory monetization).

N-Gram High Frequency Trading

Statistical analysis of financial markets Stack: Python, Pandas, Statistical Mathematics.

The Pitch: In 2021, I wanted to test whether a classical statistical approach (N-Grams, often used in linguistics) could predict micro-variations in the cryptocurrency market, without using 'black box' Machine Learning.

The Technical Analysis: I developed a bot capable of analyzing high-frequency price sequences.

Performance: The model achieved a 'Hit Rate' (success rate) of approximately 53%.
Optimization: The code had to be extremely fast to react to fluctuations.

The Economic Discovery: This project highlighted a mathematical structural inequality:

With 'VIP' transaction fees (large capitals) at 0.005%, the model is profitable.
With 'Standard' fees (small capitals) at 0.01%, the model becomes losing (effective Hit Rate of 48% after fees). This project demonstrates that profitability in markets is not only related to risk-taking, but to the entry barrier imposed by fees. I stopped the project, refusing to participate in what I consider 'dead energy', but the code remains solid proof of my algorithmic skills.