Aditi Kumari

I built NeuroStage DSS, a full-stack clinical decision support system for Alzheimer’s disease staging, with a Dockerized FastAPI backend deployed on Render and a React frontend deployed on Vercel.

For NeuroStage DSS, I designed a dual-CNN architecture with attention layers for MRI feature extraction and a soft-voting ensemble that classified four disease stages with 99.3% accuracy. I tuned the models with GridSearchCV and used SMOTE to address class imbalance.

I also built StudyLens, an AI-powered study planning platform that turns a topic into a five-day plan, summaries, flashcards, quizzes, and a spaced-repetition schedule. It uses a single-prompt generation pipeline, and I added guest access with data that persists when a user later signs up.

During my Green Intern program with 1M1B Green Skills Academy, in collaboration with AICTE and supported by Salesforce and SEMI, I built a React carbon footprint calculator as a live project deliverable. I also founded and served as the first President of my college’s Coding Club, organizing a five-day peer-mentoring program in C programming.