PharmAI - Smart Pharmacy Assistant

An AI-driven web platform to support pharmacists with stock forecasting, skin diagnostics, prescription understanding, and interactive Q&A—built with cutting-edge models and a sleek interface.

---

🚀 Overview

Developed at ESPRIT School of Engineering, PharmAI leverages state-of-the-art AI and full-stack web technologies to enhance pharmacy operations and improve patient experience.

---

✨ Features

• 📦 Stock Forecasting — XGBoost-based AI predicts medication demand trends
• 🧴 Skin Analysis — YOLOv8 detects imperfections, while EfficientNet-B0 classifies skin types for personalized product suggestions
• 💊 Prescription Understanding — Uses LLaVA for image-to-text and summarization
• 🔍 Drug Detection — Custom CNN-based OCR reads drug names from package images
• 💬 Chatbot Assistant — Agentic RAG-powered assistant handles medication questions, side effects, dosage, and symptom guidance
• 🌐 Modern UI — Responsive web interface powered by Next.js & Tailwind CSS
• 🧪 Validated by Experts — Tested in collaboration with a licensed pharmacist

---

🧠 Tech Stack

Frontend

• ⚛️ Next.js
• 🎨 Tailwind CSS

Backend

• 🚀 FastAPI
• 🐍 Python: Pandas, Scikit-learn, OpenCV

AI Models

• 🧠 YOLOv8 (Object Detection)
• 📊 XGBoost (Time-Series Forecasting)
• 🔤 Custom CNN OCR
• 🧑‍⚕️ Agentic RAG Chatbot
• 🖼️ LLaVA (Multimodal Vision-Language)
• 🧬 EfficientNet B0 (Skin Classification)

DevOps

• 🐳 Docker
• 🚂 Railway

---

🛠️ Getting Started

1. Clone the repository

   git clone https://github.com/mazighahmed/AI_Project_pharmAI.git
   cd AI_Project_pharmAI

2. Set up the frontend

   cd frontend
   npm install
   npm run dev

3. Set up the backend

   cd backend
   pip install -r requirements.txt
   python backend.py

Acknowledgments

Created by students at Esprit School of Engineering.
Special thanks to our pharmacy partner, Mr. Makram Frigui, for testing the platform and providing valuable domain-specific feedback. Gratitude also to Mr. Razi Miliani for his guidance throughout the project.