MediSense
Description
A unified health diagnostic platform that integrates four powerful machine learning models—Detect Sepsis, Diabetes, Stress, and Brain Tumor—all in one place. Medi-Sense aims to assist users in understanding potential health risks through AI-driven predictions.
Technologies
Frontend: React.js, TailwindCSS
Backend: Node.js, Express.js, MongoDB
ML: Python, Scikit-learn, TensorFlow, Flask
Medi-Sense is a hackathon project built to bring multiple healthcare prediction models into a single, unified platform. It combines features like brain tumor detection from MRI scans, stress level analysis, diabetes risk prediction, and sepsis detection—wrapped in a simple, responsive web interface powered by a Node.js and Flask backend.
Sepsis is a critical condition caused by the body's extreme response to infection, and early detection can make the difference between recovery and severe complications. It's often difficult to identify in its early stages, which makes predictive systems valuable in highlighting potential risk before it escalates.
I worked on the sepsis prediction module, using publicly available datasets to train and evaluate different models, including linear approaches and random forests. Within the limited time of the hackathon, I focused on building a working pipeline and achieved around 82-90% accuracy across variations, contributing to the overall goal of making the platform practical and usable.