Projects

Software Development

Automatic Medical Database (AutoDB), Duke Department of Computer Science

Skills: Python (SQLite, DuckDB, Langchain & OpenAI API), Google BigQuery, Javascript (React, Observable Plot)

• Created an end-to-end AI application to facilitate electronic health records data analytics for healthcare researchers.
• Utilized OpenAI GPT-4 API to translate natural language prompts to SQL and visualization code, and boost model performance to 419% by implementing state-of-the-art few-shot and chain-of-thought prompting methods.
• Designed and developed GUI with JavaScript React, and Observable Plot libraries.
• Presented results in the Duke CS+ summer research program assembly meeting and poster session (Jul. 2024); Tested on the Duke Medical School Datathlon and received unanimous positive feedback. (Apr. 2025)

PyPI IEEG package, Cogan Lab

Skills: Python, C, Cython; YAML

• Revamped intracranial electroencephalography data processing pipelines from Python to C to enhance efficiency.
• Led the packaging and publishing of the pipelines as the IEEG package on PyPI.

RAvolution, COMPSCI 316: Intro to Database Systems @ Duke

Skills: PostgreSQL (Prisma ORM), Javascript (React, Next.js), HTML/CSS

• Built a database-backed web application designed for Duke’s Housing & Resdiential Life office, offering modules for account management, activity organization, safety reporting, scheduling, and more.
• Implemented the user account module, with functionalities of registering accounts with role differentiation (resident, RA, RC), as well as searching user profiles with keyword filters.