I’m Zhuyiheng Chu, a Master’s student in Computational and Applied Mathematics at the University of Chicago. I hold a Bachelor’s degree in Information and Computing Science from Zhejiang University.

I have a passion for solving complex problems, as demonstrated by my participation in numerous programming and mathematical competitions.

Competitions & Awards

• Finalist, 2023 Mathematical Contest In Modeling (Top 1%)
• Second Prize, 2022 Contemporary Undergraduate Mathematical Contest in Modeling (Top 2%)
• Gold Medal, 2021 Collegiate Computer System & Programming Contest, East China Site
• Gold Medal, 2021 China Collegiate Programming Contest, Guangzhou Site
• Silver Medal, 2021 ICPC Asia Shanghai Regional Contest
• Gold Medal, 2020 ICPC Asia Nanjing Regional Contest
• Gold Medal, 2020 China Collegiate Programming Contest, Mianyang Site

I’ve also gained industry experience during my internship at Momenta, where I applied deep learning algorithms to predict optimal vehicle paths using multimodal data. Currently, I’m working as a grader for the University of Chicago, evaluating Python algorithm design assignments for a Master’s course.

Internship Experience

Weride (San Jose, California)

Software Engineer Intern June. 2025 – Sept. 2025

• Developed and optimized reinforcement learning models for smart agents, improving the adaptability and decision-making of the simulation system.

University of Chicago

MPCS Grader&TA | Sept. 2024 – Present

• Course: MPCS 55001 Algorithm, MPCS 55005 Advanced Algorithms
• Graded Python-based algorithm design assignments and contributed to the development of coursework for the University of Chicago’s Master of Computer Science program.
• Organized office hours to support students’ learning.

Momenta (Hangzhou City, China)

Deep Learning Algorithms Engineer Intern | Jan. 2024 – May 2024

• Using deep learning algorithms to process multimodal information perceived by the vehicle to predict the most appropriate vehicle path.
• Employing C++ to process the output results of the multimodal model to make the results more stable.

Research Experience

Reinforcement Learning for Biological Neuronal Networks

AI+Science UChicago Hackathon Apr. 2025

• Replaced traditional PCA/static state encodings with Transformer-based representations in a reinforcement learning framework, enabling Q-learning to outperform all competitors and secure 1st place in the competition. DSI website