Building Scalable AI models
tasnif
A fullstack web app with guest and registered user modes. It integrates a machine learning + NLP model for Arabic text classification into 7 categories (politics, medicine, sports, etc.). Registered users can also view their test history.
Built with Next.js + Flask + MySQL database. The project was awarded 'Best Project' in Master 1 at the university and certified accordingly.
Project Info
University Project
Web Apps, Machine Learning, NLP
2024
Constantine, Algeria
QynAI
A fullstack web app with guest and registered user modes. It integrates a deep learning + NLP model for AI-text detection in paragraphs and articles, built from scratch using CNN and BLSTM neural networks. Registered users can access their test history.
Built with Next.js + FastAPI + MySQL database. This project was recognized as the best Master 1 project at the university and certified accordingly.
Project Info
University Project
Web Apps, Deep Learning, NLP
2024
Constantine, Algeria
Lifeplus
Lifeplus is an AI-powered web application for skin cancer detection. Using a Convolutional Neural Network (CNN), it distinguishes between benign and malignant skin lesions, achieving 88.6% accuracy and a macro-averaged ROC-AUC of 0.95 on the HAM10000 dataset.
The app integrates a Next.js frontend with a FastAPI backend. Users can upload skin lesion images and receive instant predictions with confidence scores. Planned improvements include multi-class classification, advanced preprocessing, hybrid architectures, and mobile/edge deployment for real-time clinical use.
Project Info
Personal Project
AI, Deep Learning, Medical Imaging, Web Apps
2025
Algeria
NepsyAI
NepsyAI is an intelligent web-based platform designed for early screening and personalized support of neurodevelopmental disorders (NDD), with a particular focus on Autism Spectrum Disorder (ASD) and Attention Deficit Hyperactivity Disorder (ADHD). It integrates validated clinical questionnaires with machine learning models to provide accurate, interpretable results.
The system combines a screening engine with an intelligent conversational assistant built using fine-tuned LLMs, RAG, and Gemini API, enabling empathetic interaction, detailed feedback, and actionable guidance. It offers multilingual support (English, French, Arabic), secure data management, and features like screening history, user dashboards, and clinic recommendations. This project highlights the potential of combining validated clinical tools with modern AI to improve accessibility and early detection in mental health.