A production-grade, goal-driven Agentic AI chat system that enables intelligent conversations by reasoning, planning, and acting through LLMs, tool integration, and memory.

This project implements a production-oriented Retrieval-Augmented Generation (RAG) system that can answer questions over multiple document types such as PDFs, DOCX, TXT, Markdown files, and web URLs.

An end-to-end NLP system that ingests raw news articles, and serves real-time predictions to classify news into categories like World, Sports, Business, and Tech. through a FastAPI interface.

An industry-grade, production-ready Machine Learning system that predicts vehicle insurance outcomes using a complete MLOps pipeline.

A deep learning-powered web application for automatic brain tumor classification from MRI images. Built with Flask, TensorFlow/Keras, and deployed on Render.

This app allows users to input patient attributes such as age, chest pain type, cholesterol, and more. The backend uses a Random Forest model to predict whether the patient is likely to have heart disease.

A computer vision system that tracks people entering and exiting through a designated door area using YOLO object detection and ReID (Re-Identification) techniques.

Researchers developed a deep learning model using MRI images to classify brain tumors into four types: Glioma, Meningioma, Pituitary, and No Tumor. The goal is to create an accurate diagnosis tool for medical experts.

The research paper aimed to develop a machine learning-based approach for early identification of HCV, which could potentially improve treatment outcomes.