Developed an end-to-end MLOps pipeline and custom Chrome Extension for real-time YouTube sentiment analysis, achieving 79.12% accuracy using a LightGBM model deployed on AWS EC2 with automated CI/CD.
“A secure, RAG-driven medical diagnosis engine featuring hybrid OCR document parsing, longitudinal trend analysis, and a doctor verification loop, powered by LLaMA 3.3, Pinecone, and FastAPI.”
Engineered an end-to-end deep learning pipeline to classify 43 German traffic signs (GTSRB) with 98.73% accuracy, deployed via a Streamlit dashboard for real-time inference and performance monitoring.
