Clinical AI

Clinical AI Platform with Symptom Analyzer is a sophisticated healthcare application developed for the IBM BoB Hackathon that combines modern web technologies with advanced machine learning to create an intelligent clinical decision-support system. The platform is built with a React 19 frontend (featuring TypeScript, React Router, and TailwindCSS for a responsive user interface) and a Flask backend powered by Python and PyTorch for deep learning capabilities. The project's centerpiece is an AI-powered symptom analyzer that leverages BioMedCLIP, a medical imaging model developed by Microsoft Research. This model has been fine-tuned using LoRA (Low-Rank Adaptation) on the HAM10000 dataset, which contains over 10,000 dermatoscopic images of skin conditions. The analyzer achieves an impressive 85-95% accuracy rate in detecting and classifying 7 different types of skin lesions. Beyond simple classification, the system provides comprehensive analysis including automatic severity assessment (mild/moderate/severe), personalized action recommendations, confidence scoring for transparency, and alternative diagnoses with confidence metrics. The platform offers comprehensive clinical features including patient profile management (tracking conditions, allergies, and medications), a post-appointment hub for clinical note uploads and summaries, an intelligent AI chat panel with emergency capabilities, and integrated task management for clinical documentation. The modular architecture supports scalable deployment with detailed setup instructions and production-ready configuration using Gunicorn. While designed for educational and demonstration purposes, the project represents a full-stack implementation of clinical AI, combining modern web development practices with state-of-the-art machine learning techniques to support healthcare professionals.