CS student Wei Xiong named Google PhD fellow

Every year, the Google Ph.D. fellowship program chooses a handful of outstanding graduate students to support them through their research journey. In 2025, Wei Xiong, a Computer Science Ph.D. student at The Grainger College of Engineering at the University of Illinois Urbana-Champaign, was awarded a spot in the program. 

​​My research focuses on how to make large language models learn and improve after they finish pre-training. This stage is often called post-training and is where models learn to follow instructions, reason more reliably, and behave in ways that are helpful and safe for people.
–Wei Xiong

In his current research, Xiong is focusing on reinforcement learning (RL), an area of machine learning where an AI system improves through trial and error. His ultimate goal is to improve the training methods for the algorithms that support LLMs. 

Throughout this process, Xiong is aiming to answer the bigger question of how to scale up the RL in large-scale models. In order to do this, he would have to figure out a mathematically sound way to fit this feature into the large models. 

“ If we can make RL as scalable and reliable as pre-training, we unlock a powerful path forward: models that continue to improve beyond the limitations of real-world datasets, learning new skills directly through interaction,” Xiong said. “Understanding how to achieve this, algorithmically and system-wise, is the central challenge that motivates my research.”

As far as his course of action goes, Xiong is determined to do two things: develop principled and mathematically grounded algorithms and work closely with Google’s industry teams. In his plan, Xiong acknowledges how necessary it will be to establish stability in all aspects of his research. 

“I want to make large-scale RL more efficient and stable,” Xiong said. “RL for LLMs cannot be fully understood without large-computer settings, so partnering with Google’s research teams will be essential.”

Despite finding motivation in his interest for improving RL and LLMs, one of the factors that persuaded Xiong to apply for the program was the good experience he had at his Google DeepMind summer program during the first year of his Ph.D. The mentorship and environment had a major influence on why he wanted to pursue another opportunity studying AI with Google. 

“Applying for the Google Ph.D. Fellowship felt like a natural way to continue that collaboration and contribute to the broader mission of advancing responsible AI,” Xiong said.

The Google Fellowship supports Wei Xiong's research; he works with CS professors Tong Zhang and Nan Jiang, and he is a researcher at OpenAI.

Grainger Engineering Affiliations

Tong Zhang is an Illinois Grainger Engineering professor of computer science and is affiliated with the Siebel School of Computing and Data Science.

Nan Jiang is an Illinois Grainger Engineering associate professor of computer science and is affiliated with the Siebel School of Computing and Data Science.