Iโm Camil an EPITA last year student, and Iโve got a habit of diving into projects that are half science, half experiment. My current obsession? Trying to read numbers from EEG signals. Itโs a mix of weirdness and discovery that keeps me hooked, and I wouldnโt have it any other way. Hackathons are where I feel at homeโsomething about building wild ideas in no time just clicks with me
Currently building ...
- Description: Developing a model to predict handwritten digit classifications from EEG signals, effectively combining neural data with classic MNIST image recognition for an innovative approach to mind-to-number interpretation.
- Technologies:
-
Description: Echo helps support loved ones with Alzheimer's by creating a shared album. Family members can post souvenirs, pictures, and comments. Echo features an AI-powered chatbot for retrieving past posts and personalized memory quizzes using ๐ค๐ (face recognition) and ๐ฒ Thompson sampling, ensuring an adaptive quiz experience.
-
How We Built It: Built with React (frontend) and FastAPI (backend). The app has three main interfaces for data handling, chatbot interactions, and quiz generation using Thompson sampling to balance exploration and exploitation for better memory retention.
- Description: Created a decentralized authentication system using Fully Homomorphic Encryption (FHE) to secure biometric data, ensuring data remains encrypted end-to-end.
- Technologies:
- Description: Developed a platform for converting large pre-trained models into efficient, task-specific smaller models using prompt engineering and direct optimization.
- Technologies:
- Description: Created a multi-crypto trading bot for memecoins using blockchain big data with a micro architecture for daily model updates.
- Technologies:
- Duration: September 2024 - January 2025
- Description: Studied tumor heterogeneity using MRI-guided MRIDIAN imaging and a radiomics approach for survival analysis based on patient clusters.
- Technologies:
- Duration: September 2023 - January 2024
- Description: Developed a RAG/LLM agent system using LangChain and Weaviate for multi-modal content generation.
- Technologies: