Richard Zhu studying different aspects of machine learning, from natural language processing to analyzing fluid dynamics

April 3, 2023

Richard Zhu, age 21, Class of 2023


Zhu is a senior in the mechanical and aerospace engineering department. He is pursuing three undergraduate certificates, one from the Center for Statistics and Machine Learning (CSML), the second from the Program in Applied and Computational Mathematics, and the third from the Program in Applications of Computing. He is also a CSML undergraduate student ambassador.

Undergraduate Work:

Zhu has been a research assistant for Princeton Natural Language Processing, a lab led by professors Sanjeev Arora, the Charles C. Fitzmorris Professor in Computer Science, and Danqi Chen and Karthik Narasimhan, both assistant professors of computer science.

His work in the group involves creating bots that play Diplomacy, a Risk-like game played often on a pre-World War I map of Europe. Unlike Risk, players can also issue press releases, allowing them to communicate via text to form alliances and announce hostilities. For such bots to perform well, they must excel at detecting deception in the virtual environment of the game. Denis Peskoff, a Princeton postdoc, wrote about deception detection in the game in “It Takes Two to Lie: One to Lie and One to Listen,” which was published at the Association for Computational Linguistics conference in 2020. Zhu is building upon this work by designing and training a generalized reinforcement learning model that allows accurate measurement of each competing bot’s intentions.

For his senior thesis, a part of which will serve as his CSML independent project, Zhu is studying droplet formation from coaxial nozzles, a specialized nozzle in which two different types of fluids or materials are extruded and combined into one stream.

“Atomization of fluid impacts our daily lives in unseen ways, with atomizers being used to apply everything from Insecticides to car paint. This work investigates the parameters for droplet formation within coaxial nozzles by studying the droplet distribution of fluids as they leave sub-millimeter nozzle openings,” said Zhu.

The first part of his project will involve the fabrication of the nozzle out of resin using a high-resolution 3D printer. Then Zhu will use various data science techniques to process and analyze the statistical distribution of particles from the nozzle. Data includes droplet diameter, consistency of droplet size, and the ejection pattern for particles in suspension. He plans on developing both theoretical and empirical models to accurately describe particle distribution.

“Understanding the properties of the resulting flow has implications for both novel aircraft materials and for jet or rocket engine fuel injectors,” Zhu said, whose thesis advisor is Howard Stone, the Donald R. Dixon '69 and Elizabeth W. Dixon Professor of Mechanical and Aerospace Engineering.

Outside of his undergraduate studies, Zhu has served as an investment banking summer analyst for Citi and TD Securities. He has also been a software engineering intern at Meta and nanopay, a global payments technology company.

After graduation, Zhu is planning to attend the master’s in engineering program at Princeton’s mechanical and aerospace engineering department. Zhu hopes to spend this time exploring graduate course material and continuing his undergraduate research interests before returning to Citi.

Extracurricular activities:

Zhu has been involved in Manna Christian Fellowship on campus, served as orientation leader for Outdoor Action for three successive years, and is a member of the Cap and Gown Club. He has also been a team member of Princeton’s College Fed Challenge team, in which undergraduates “analyze economic and financial conditions and formulate a monetary policy recommendation, modeling the Federal Open Market Committee,” according to the competition website, placing first in the 2022 Challenge.

For Fun:

Zhu enjoys running, reading and cooking in his down time.