I’m a junior at the University of Rochester majoring in Computer Science and Brain & Cognitive Sciences. My interests lie at the intersection of Artificial Intelligence and Human–Computer Interaction, where I explore how intelligent systems can enhance human creativity, learning, and mixed reality experiences. I’m passionate about designing technologies that bridge computational reasoning with human-centered design to create more intuitive and adaptive interactive systems.

Projects

Mancala

An AI-based Mancala agent that evaluates moves using heuristics and game-tree search. This project implements an intelligent agent that plays Mancala using search-based strategies and evaluation functions. It deepened my understanding of game trees, minimax search, and heuristic design in adversarial AI environments.

Sudoku Solver

A Constraint Satisfaction Problem (CSP) solver for Sudoku puzzles. This project explores constraint satisfaction, a foundational concept in AI. It models Sudoku as a CSP with variables, domains, and constraints, finding efficient solutions even for an NP-complete problem. This work enhanced my understanding of AI search methods used in scheduling, planning, and optimization.

Polyhunt

A hands-on project connecting the mathematics of linear regression to practical implementation. I implemented both a closed-form solver and a gradient descent solver, explored hyperparameters and regularization, and analyzed trade-offs between analytic and iterative optimization approaches in machine learning.

Research Experience

BEAR Lab — University of Rochester

• Work with Prof. Yukang Yan and Ph.D. students on developing and testing Mixed Reality prototypes.
• Build Unity-based Mixed Reality demos for ongoing research projects in the BEAR Lab.
• Explore environmental features and visual elements to enhance Mixed Reality realism.

Computational Physicality Lab — National Taiwan University

• Collaborated with Prof. Lung-Pan Cheng and graduate researchers on Puzzle Me This, a Mixed Reality system that generates everyday object puzzles.
• Focused on Unity scene construction and room/object scanning workflows for Mixed Reality prototypes.
• Collected scans and evaluated multiple SLAM pipelines (e.g., DROID-SLAM / related modern variants) and pose-estimation approaches (e.g., FoundationPose / OnePose++ style setups) to compare tracking stability, drift, and runtime.