Crafting AI Solutions & Architecture
Data Scientist and AI Engineer specializing in AI/ML innovations, GenAI/AI Agents, and AI research for Healthcare
Background
As a Data Scientist, AI Engineer, I am deeply involved in the development and deployment of advanced AI/ML solutions. My work combines cutting-edge research in healthcare AI, GenAI and AI agents, with hands-on experience in building practical applications across diverse domains. I leverage my expertise in machine learning, deep learning, and natural language processing (NLP) to create impactful solutions that address real-world challenges.
AI & Machine Learning
Cloud Architecture & MLOps
Programming
Data Science
Featured Projects
A selection of my recent work in cloud architecture and full-stack development
SportSphere AI Agent
A conversational AI agent/assistant for sports enthusiasts, providing real-time information on MLB, NBA and Global Soccer. Interact via voice or text to get live scores, team rosters, player statistics, and league standings.
Pocket Traveler
Pocket Traveler is an AI-powered personal travel assistant that creates fully customized itineraries based on your preferences. Input your travel details (origin, destination, dates, and more) via text or voice, and Pocket Traveler handles the rest, generating personalized flight options, accommodation suggestions, activities, dining recommendations, local insights, and even safety tips. The complete itinerary is delivered directly to your inbox.
CustomGPT
CustomGPT is a versatile LLM chatbot built with Streamlit that integrates multiple language models (OpenAI, DeepSeek, HuggingFace, and Ollama) and implements Retrieval-Augmented Generation (RAG) techniques (both Simple RAG and LightRAG). It supports various RAG operations, including document uploads, web link analysis, GitHub repository context, and research paper integration. It utilizes Pinecone and Chroma for vector databases and stores chat history in MongoDB.
License Plate Detection
A Streamlit-powered application for license plate detection using YOLOv8. The app allows users to upload images or videos, and it detects and highlights license plates within the provided media. The training data was sourced from Roboflow.