research
Single-cell transcriptome of retinal myeloid cells in response to transplantation of human neurons reveals reversibility of microglial activation
Preprint on bioRxiv
Paper publishing to Molecular Therapy
Poster accepted to 2025 Children's Tumor Foundation NF Conference
This research investigates the cellular responses of retinal myeloid cells following human neuron transplantation, using single-cell RNA sequencing to reveal the dynamic and reversible nature of microglial activation in the context of retinal regenerative therapies.
A Comprehensive Exploration of Digital Twinning in Spine Surgery
Published in Clinical Spine Surgery
Poster presented at Congress of Neurological Surgeons Conference, Houston, TX. October 2024
This study explores the application of digital twin technology in spine surgery, examining how virtual replicas of patients can enhance surgical planning, improve outcomes, and reduce complications through advanced modeling and simulation techniques.
Poster presented at TJHSST Science and Engineering Fair
This research develops a comprehensive single-cell RNA sequencing atlas of mouse retinal cells across developmental stages (E11 to P14), providing researchers with a platform for investigating changes in intrinsic retinal genetic heterogeneity and extrinsic cell-cell interactions over time to improve retinal organoid modeling for disease research.
Poster presented at 2025 ASSIP Poster Fair
Conducted at the Aspiring Scientists Summer Internship Program, Mobility Observatory and Data Analytics Lab, of George Mason University
This project applies a spatial mismatch framework to quantify where heat hazard exceeds cooling access in New York City, integrating FEMA NRI metrics, cooling-access measures, and demographic data to identify high-risk, low-access clusters and produce actionable targets for equitable cooling center placement.
Evaluating AI-Generated Latin Narratives for Early Language Learners
Conducted at the Quantitative Criticism Lab of Harvard University, The University of Texas at Austin
This IRB-submitted study uses large language models to generate Latin-adapted superhero stories for early learners, addressing gaps in accessible pedagogy materials. Methods include prompt engineering with n-shot examples aligned to Junior Classical League standards and mixed-methods evaluation with in-class studies.
The Effect of Synthesis Method and Starting Materials on the Yield and Viability of Quantum Dots
Conducted at TJHSST's Chemical Analysis & Nanochemistry Lab
This study synthesizes carbon quantum dots from lemon juice and polyethylenimine using a one-pot heating method. Unlike semiconductor quantum dots with toxic heavy metals, these carbon-based nanoparticles offer strong optical performance for water and air quality monitoring applications.