Comcast

During my internship, I successfully increased operational efficiency by 30% through the development of a real-time analytics dashboard and optimizing a RAG model, improving response accuracy by 25%

Ongoing

Help use pre-snap behavior to predict and better understand NFL team and player tendencies. we are working on developing a Confusion Score—a numerical representation of how disorganized a defense appears before the snap.
github repository

This project is to analyze and determine the sentiment of videos using both visual and audio cues. Enhances video sentiment analysis by combining visual and audio cues using deep learning. Integrating facial expressions, body language, and speech improves accuracy over single-modality methods.
github repository

This project involves developing an image processing web application using Django, with AWS services integrated to enhance scalability and reliability. The primary AWS services utilized include EC2 instances to host the application, an S3 bucket for storage, and an Elastic Load Balancer (ELB) to manage incoming traffic for high availability.
github repository

The Stock Screener App is an interactive, web-based stock screening tool built with Dash. It provides users with powerful filtering, analysis, and visualization features to screen stocks effectively and make data-driven decisions.
github repository

Developed a tool that clones GitHub repositories and creates a document retrieval system using FAISS, OpenAI embeddings, and LangChain for natural language queries.
github repository

Created a clustering system using unsupervised learning to analyze global economic, social, and geopolitical trends, enabling informed decision-making.
Link to the page github repository

Developed a content-based movie recommendation system using NLP and cosine similarity, deployed as a Streamlit web app.
Link to the page github repository