Low-Resource Language Models for Kurdish and Narragansett

Developed low-resource language models for Kurdish and Narragansett using BERT and HuggingFace in Python, contributing to more inclusive natural language processing technology for 1,800+ people in underserved linguistic communities across Rhode Island. Focused on equitable language model training, bias detection, and linguistic preservation.

Urban Air Quality Analysis in Chinatown

Built a Python tool integrating web scraping and generative AI to streamline literature reviews on PubMed, reducing manual search time by 25%. Analyzed urban air quality disparities in disadvantaged Boston neighborhoods using Python, R, and Excel. Presented findings to 200+ community members at Tufts Medical Center and Boston Chinatown Neighborhood Center to inform advocacy efforts.

Direct Air Capture via Metal-Organic Supercontainers

Conducted experiments in direct air carbon capture and drug delivery. Applied computational and statistical tools for data interpretation in Python, R, and Microsoft Excel. Conducted literature reviews, authored and edited grant proposals, graduate-level theses, and research papers. Presented findings at American Chemical Society Regional Conferences (2022, 2023).