Research
My work focuses on computational metascience: building data-driven tools to understand and improve how scientific research is done. Iām especially interested in collaboration systems, recommender algorithms, and research infrastructure that supports transparency, reproducibility, and discovery.
Current Research
Measuring Research Impact with NLP
This project uses natural language processing to study how scientific research produces impact. I analyze the language of publications, collaboration patterns, and institutional context to trace how ideas develop, spread across disciplines, and lead to meaningful outcomes. The goal is to build computational tools that help researchers and institutions better understand what structures and practices lead to effective, high-impact research.
Past Research
CABPortal: Research Collaboration Hub
I rebuilt CABPortal from the ground up using Python, Flask, and SQL to serve as a central platform for interdisciplinary research collaboration. The portal enables faculty and students to connect, track interests, and recommend matches through an integrated recommender engine.
CAB Intelligence: NLP-Based Recommender
As the intelligent engine behind CABPortal, CAB Intelligence is a natural language recommender system I designed to suggest interdisciplinary collaborators and projects. It uses text embeddings, indirect signals, and synthetic prompts to improve matching in low-data environments.
Post-Starburst Galaxy Simulations
Created open-source scripts to simulate and analyze telescope observations of post-starburst galaxies. This work helped constrain evolutionary parameters by modeling how galaxies appear at different stages.
Groundwater Subsidence Modeling
Used InSAR and high-performance computing to model ground collapse due to aquifer depletion in New Jersey. Also explored planetary analogs by comparing patterns with similar features on Mars and the Moon.
Number Theory & HPC
Built high-performance computing tools to explore properties of Euclidean and norm-Euclidean quadratic number fields. Focused on optimizing algorithms for real and imaginary domain analysis.