Imagination Trumps Knowledge!
Featured ProjectReimagining Walmart's Shoping Experience
Built an intelligent AI shopping agent using LangGraph that turns natural language recipes into personalized grocery lists. Integrated with a Walmart-clone platform, the agent fetches relevant ingredients, suggests brands, and adds user-preferred items to the cart. Powered by Supabase, FastAPI, WebSockets, and a modern Next.js + React frontend.
ProjectEngineered an MCP Server for Apple Reminders using Python and FastMCP, enabling seamless, protocol-based interaction with the native macOS Reminders app. Combined the power of AppleScript automation with backend intelligence to create a smooth, end-to-end integration. Think of it as giving your Reminders app a programmable brain—automated, tested, and built for developersApple Reminders MCP Server (MacOS Application)
ProjectA context-aware Retrieval-Augmented Generation (RAG) system designed to enhance response accuracy by intelligently retrieving only the most relevant information based on the context of user queries. The system incorporated metadata tagging to effectively categorize and organize documents, enabling efficient retrieval of specific content chunks. This approach significantly improved performance by reducing noise and ensuring that the generation model received high-quality, contextually appropriate data for response generation.Contextual Retrieval RAG Bot
Research Paper Published in IEEEHybrid Quantum Model for Digital Media Processing for Metal Surface Defect
Leveraged quantum computing techniques within the PennyLane framework to improve digital image analysis for crack detection on metal surfaces. Developed a hybrid model by integrating Convolutional Neural Networks (CNNs) with quantum circuits, aiming to enhance pattern recognition and classification accuracy. Utilized an open-source dataset for training and evaluating the model, demonstrating the potential of quantum-classical approaches in non-destructive testing and industrial inspection applications.
ProjectDeveloped a KNN-based Food Recommendation System trained on a public dataset comprising 48,735 entries and 12 user dietary features, using TfidfVectorizer for effective feature extraction to deliver personalized food suggestions based on individual preferences. To enhance the recommendation quality and variety, I integrated the Spoonacular API, providing access to over 600,000 products, 5,000+ recipes, and 115,000+ menu items, enriched with detailed data on nutrition, pricing, and cooking tips. The system was built with a Flask backend and an HTML/CSS frontend, enabling users to receive real-time, customized food recommendations.Food Recommendation System
ProjectDuring a hackathon, I developed a Diabetes Prediction Model using Artificial Neural Networks (ANN), achieving 85% accuracy on a public dataset containing 1,000 records and 13 key health indicators such as insulin levels, glucose, and blood pressure. To make the model accessible and user-friendly, I built a full-stack web application featuring a React + JSX frontend and a Flask backend. This application allows users to input real-time health data and receive instant diabetes risk predictions, supporting early assessment and proactive health management.Early Stage Diabetes Prediction