An undergraduate researcher at Princeton University studying Neuroscience — with research across educational measurement and generative-AI detection (University of Maryland), computational genomics (Westlake University), and computational social science (Princeton University).
Supporting interdisciplinary research using machine learning and multimodal data to study human behavior and communication on social media — contributing annotated training data and identifying behavioral, linguistic, and contextual features relevant to computational analyses of online discourse and social interaction.
Research at the intersection of computational genomics, machine learning, and deep learning — with an emphasis on genomic sequence modeling, molecular phenotype prediction, model generalization, and rigorous benchmarking. Developing and evaluating deep-learning pipelines for biological data with large-scale datasets and GPU-based computational research environments.
Research on automated essay scoring, AI-generated text detection, and educational measurement using machine learning, natural language processing, and psychometric methods — developing transformer- and LLM-based approaches, text summarization, linguistic features, and ensemble methods, with peer-reviewed publications and presentations at IMPS, NCME, and other international conferences. This includes the automated scoring system for AP humanities writing (history, literature, foreign languages), long-form essay classification, and cost-efficient generative-AI summarization for scalable scoring.
Official note taker for Professor Christiane Fellbaum’s freshman seminar Evolution of Human Language — producing structured and accessible academic notes that synthesize linguistic theories, empirical findings, and classroom discussions.
Building an automated scoring system for AP humanities essays — history, literature, and foreign languages. The goal is a scoring pipeline that is accurate, transparent, and equitable: refining evaluation criteria, exploring resource-efficient modeling, and expanding applicability across prompts and contexts.
At Princeton University: machine learning on multimodal data to study human behavior and communication on social media — contributing annotated training data and identifying behavioral, linguistic, and contextual features for computational analyses of online discourse and social interaction.
At Westlake University: developing and evaluating deep-learning pipelines for biological data — genomic sequence modeling, molecular phenotype prediction, model generalization, and rigorous benchmarking — with large-scale datasets and GPU-based computational research environments.
Showed that generative-AI summarization can make automated essay scoring scalable and cost-efficient in educational assessment, reducing inference cost while preserving scoring quality.
Compared scoring rationales between large language models and human raters, explored how summarization by generative models supports scoring of long essays, and examined the utility of LLM rationales for enhancing automated essay scoring.
Studied the impacts of tokenization and dataset size on identifying AI-generated text (Frontiers in Artificial Intelligence), and comparatively evaluated machine-learning and LLM approaches for detecting ChatGPT-written essays under revision conditions (IMPS 2024 Proceedings).
Machine learning and multimodal data for human behavior and communication on social media; annotated training data and feature identification for computational social science.
Deep-learning pipelines for genomic sequence modeling, molecular phenotype prediction, and rigorous benchmarking in a GPU-based computational research environment.
Automated essay scoring, LLM scoring rationales, and generative-AI text detection for K–12 and AP assessment contexts, with Prof. Hong Jiao’s group.
Official note taker for Professor Christiane Fellbaum’s seminar Evolution of Human Language — structured academic notes synthesizing linguistic theories, empirical findings, and classroom discussions.
Member of the Culver Class of 2025 Alumni Leadership Team — strengthening connections among classmates and supporting alumni communication, class traditions, and continued community engagement.
Led a cadet unit of approximately 54 students, with responsibility for unit organization, accountability, mentorship, and standards.
Represented and helped lead a class of approximately 220 students — supporting class initiatives, community activities, and student engagement.
Supported the school’s Honor Code and participated in student-led efforts promoting integrity and accountability.
Senior cadet leadership involving unit organization, mentorship, accountability, and daily operations.
Led student efforts related to institutional traditions, unit history, values, and standards.
Early cadet leadership focused on military standards, peer support, and unit responsibilities.