Research Experiences

2026

MIT SERC Scholar

LLM-Based Simulation of Multilateral Climate Negotiations: A Multi-Agent Framework for Computational Climate Diplomacy

Postdoctoral mentor: Dr. Anna Papp

Explores whether LLM-based multi-agent systems can meaningfully simulate multilateral climate negotiations. The project studies how heterogeneous agents, institutional constraints, and iterative dialogue shape negotiation dynamics, coalition formation, and final agreement outcomes in settings inspired by real-world climate diplomacy.

Methods & themes: Multi-Agent LLM Systems · Computational Social Simulation · Climate Negotiation · Coalition Dynamics

GitHub Poster SERC Scholars Program

2025

Undergraduate Thesis Research

Climate Uncertainty Propagation through Statistical Downscaling and Building Energy Models

Advisors: Prof. Michael T. Craig and Prof. Xi Lu

Examined how uncertainty in future climate conditions propagates through statistical downscaling and building-energy models into residential electricity demand projections. The work generated high-resolution weather inputs for U.S. counties and linked them to county-level heating and cooling demand estimates, with a focus on long-horizon uncertainty propagation and the transition toward cooling-dominated demand.

Methods & themes: Statistical Downscaling · Climate Ensembles · Building Energy Modeling · Uncertainty Propagation

Poster Dataset Related preprint

2024

Undergraduate Visiting Research Program

Gaussian Process Regression for Predicting Precipitation with Oscillations and Reducing Uncertainty Over Time

Advisor: Prof. Sarah Fletcher

Developed Gaussian process models for precipitation projections over the Awash Basin in Ethiopia using outputs from multiple CMIP6 models. The project focused on capturing nonstationary precipitation variability by incorporating both long-term climate change signals and oscillatory behavior associated with large-scale climate modes.

Methods & themes: Gaussian Processes · CMIP6 Projections · Climate Oscillations · Precipitation Uncertainty

GitHub Poster

2023–24

Undergraduate Researcher

Renewable Resource Droughts and Power System Implications

Advisor: Prof. K. Max Zhang

Characterized periods of simultaneous low wind and solar availability in New York State and connected renewable-resource climatology with power-system adequacy. The study combined renewable-generation modeling, electricity-demand modeling, and supply–demand adequacy metrics to evaluate the duration, severity, and spatial distribution of renewable-resource droughts and their implications for deeply decarbonized electricity systems.

Methods & themes: Renewable Resource Droughts · Demand Modeling · Supply–Demand Adequacy · Power-System Modeling

2022–23

Group Leader, Student Research Training Program

Modeling and Promoting Express Packaging Reuse Behavior among Beijing Resident

Advisor: Prof. Zongguo Wen

Developed an end-user database describing the spatial and temporal distribution of express-packaging waste generation in Beijing, supported by questionnaire and interview data from 497 participants. Behavioral modeling was used to identify drivers of residents' sorting and recycling decisions, while computer-vision methods were developed to identify and categorize packaging materials from collected and online images. The project connected waste characterization, behavioral analysis, and automated material recognition to inform packaging-reuse and recycling strategies.

Methods & themes: Waste Characterization · Behavioral Modeling · Computer Vision · Circular Economy

Award: 2nd Prize (Top 5%), National University Student Social Practice and Science Contest on Energy Saving and Emission Reduction.

Poster