Featured Work
Architected full-stack podcast analysis platform integrating RSS feed parsing, speech-to-text transcription, and real-time contextual source discovery, processing audio content into URL sources.
Engineered LangGraph AI workflow orchestrating grok-4-fast to autonomously search arXiv, Wikipedia, Congress.gov, and other web sources, surfacing relevant articles and citations synchronized to podcast timestamps.
Built responsive Next.js frontend with FastAPI backend enabling real-time transcript analysis and contextual source linking, delivering seamless user experience for research/fact-checking during podcast consumption.
Youtube Transcript Chatbot★
June 2024
Engineered a LangChain Retrieval-Augmented Generation (RAG) pipeline to extract YouTube transcript data and populate a Pinecone vector database for relevant, context-aware data retrieval.
Built and deployed a NextJS and Tailwind CSS chatbot leveraging the RAG pipeline and GPT-4 Mini to provide video-specific responses.
Other Projects
Built a secure JWT-based authentication system for login/signup, ensuring robust user access control.
Created protected API routes with token verification middleware, securing user-specific features.
Developed RESTful API endpoints with CRUD functionality to power a community forum, driving mental health and financial discussions for college students.
Built middleware to handle forum post moderation and user permissions.
Integrated GPT-3.5 Turbo with custom prompts to deploy a chatbot delivering UF-specific mental health resources.
Worked with teammates to design UI/UX in Figma, keeping mobile and desktop in mind.
Tested API endpoints with Postman and documented backend processes for team efficiency.
Led a team of four in an online hackathon by coordinating tasks and making key decisions, resulting in our team reaching top 5 out of 100 teams
Created the front-end using Tailwind CSS and the Shadcn UI library, resulting in a clean and accessible user interface.
Utilized OpenAI API to develop a multi-modal agent capable of analyzing text, images, and audio for customer support complaint detection.
Leader of a team of three, organizing weekly meetings and delegating tasks, facilitating on-time project delivery.
Evaluated a dataset of 7.7 million crashes, resulting in the identification of high-risk locations
Created a hash map data structure in C++ for efficient data storage
Created as a final project for COP3530 where we were tasked to compare two data structures.