Hi, I’m @chloedia, but my real name is Chloé Daems ! I’m interested in Machine and Deep Learning applied to image, Natural Language Processing and Reinforcement learning.
In my GitHub, you can find some of my university projects and personal ones I worked on in my free time. I finished my engineering degree at CentraleSupélec and am currently working on open source RAG improvement with Quivr.
I'm currently learning Rust 🦀 — follow my journey here
I am passionate about technology, travels and art !
Don't hesitate to hit me up if you have questions on one of my project or on my profil. 💻
▶️ MegaParse
MegaParse is a high-performance, modular parser that extracts content from a wide range of document formats (PDF, DOCX, TXT, etc.) while minimizing information loss. It is part of the Quivr RAG ecosystem and is designed to preserve structure and semantics for downstream vectorization and search.
A plugin for Obsidian.md that automatically transforms your markdown notes into spaced repetition flashcards. It supports smart parsing of headers, question-answer formatting, and export to CSV for use in quiz apps like Anki.
▶️ Connect 4
A Connect 4 AI built from scratch using the Minimax algorithm and alpha-beta pruning for performance. It includes a custom heuristic evaluation function to optimize play strategy against human opponents.
▶️ PacMan
A reinforcement learning comparison project using a simplified PacMan environment, where I implemented and compared Q-learning and SARSA agents to analyze their performance over time in a dynamic setting.
