A showcase of my technical, research, and creative work.
I developed internal HR chatboot tools using AWS services (Aurora, EC2, SNS), PostgreSQL, and LangChain for Generative AI use cases. I built both a Retrieval-Augmented Generation (RAG) system and an SQL Agent. The RAG system converted Excel files into vectorized documents using ChromaDB, which enabled HR to query salary-related data through natural language. The SQL Agent was also built using LangChain, AWS Aurora, Pandas, and NumPy for query and preprocessing HR data, hwile also integrating safeguards with AWS SNS for sensitive natural language inquiries. I configured AWS Virtual Private Cloud (VPC) and Aurora/RDS settings through the AWS Console while connecting to the server using pgAdmin.
This project was a full-stack web application that takes user input through a questionaire and generates a recipe as a result. This project was developed during the VTHacks 12 Hackathon. We implemented propelAuth to serve as the login/sign-up User Interface. Th efrontend web pages were built using Bootstrap and Svelte. We deployed Meta Llama on Docker to generate the recipe after the questionaire, and we implemented MongoDB to store users' previously generated recipes.
This project was developed to be an app to record a user's record times for sports. MongoDB was implemented to store users' record times. Vite.js and React Router were utilized to optimize and tie routes together. Svelte.js and Node.js were used as the main frontend and backend frameworks for this project.
I collaborated with a team of four to create a Java-based application that leveraged advanced object-oriented programming rpinciples and abstract data structures to develop an effective solution for data parsing and visualization. I delivered a tool that efficently rpocesses input data, applies insertion and selection sorting algorithms, and presents the results in a bar graph format through a GUI. I also developed robust data parsing techniques and implemented error-handling mechanisms to improve the system reliability. My work in the project, specifically, boosted the application's performance and usability with accurate data visualization and dependable sorting outcomes.