Zhengyang Geng
Zhengyang Geng is an artificial intelligence researcher who worked on deep equilibrium models and one-step generative models during his PhD studies at Carnegie Mellon UniversityCarnegie Mellon UniversityCarnegie Mellon University (CMU) is a private research university in Pittsburgh, Pennsylvania. It grew out of the Carnegie Technical Schools, founded with funding from Andrew Carnegie in 1900, which were renamed the Carnegie Institute of Technology in 1912 and took the present name after merging with the Mellon Institute in 1967. The university is known for computer science, robotics, engineering, and the arts, and has campuses in Qatar, Silicon Valley, and Rwanda. As of September 2025, it enrolled 16,582 students. Its current president is Farnam Jahanian. On September 30, 2026, the university announced a $3 billion gift from Ken Griffin and plans to build a new campus in Miami.Open the full entry, is the first author of the Mean Flows paper, and now works at OpenAI.
Contents8 sections
Key facts
2021Research on Attention and Matrix Decomposition
In September 2021 the paper “Is Attention Better Than Matrix Decomposition?,” with him as first author and co-authors including Zhouchen Lin, was submitted to arXiv.[1]
April 2022Deep Equilibrium Optical Flow Estimation
In April 2022 the paper “Deep Equilibrium Optical Flow Estimation,” which he co-authored with Shaojie Bai, Yash Savani and J. Zico Kolter, was submitted to arXiv.[2]
January 2024Contributing to the LLM Inference Framework Medusa
In January 2024 the paper “Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads,” on which he is a co-author along with Tianle Cai, Tri Dao and others, was submitted to arXiv.[3]
May 2025Publishing the Mean Flows Paper on One-Step Generation
In May 2025 the paper “Mean Flows for One-step Generative Modeling,” with him as first author and co-authors including J. Zico Kolter and Kaiming He, was submitted to arXiv, proposing a framework for one-step generative modeling. The paper’s first page lists his affiliation as Carnegie Mellon University and notes that the work was partly done while he was visiting MIT.[4][5]
2026Equilibrium Reasoners Paper
As of October 2026: At OpenAI
As of October 2026, his personal website describes him as a researcher at OpenAI and says he was a PhD student working with Zico Kolter and Kaiming He; it lists his research interests as adaptive computation (from one-step models to deep equilibrium models) and dynamical systems and training dynamics.[7]
Related Organizations, Websites, and Public Accounts
Carnegie Mellon University: Official website: https://www.cmu.edu/ (opens in a new window)[5]
OpenAI: Official website: https://openai.com/ (opens in a new window)[7]
Personal website: Public link: https://gsunshine.github.io/ (opens in a new window)[7]
X: Public account: https://x.com/ZhengyangGeng (opens in a new window)[7]
GitHub: Public account: https://github.com/Gsunshine (opens in a new window)[7]
Google Scholar: Public profile: https://scholar.google.com/citations?user=lNkw3QYAAAAJ (opens in a new window)[7]
Sources
- Is Attention Better Than Matrix Decomposition? (opens in a new window)
- Deep Equilibrium Optical Flow Estimation (opens in a new window)
- Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads (opens in a new window)
- Mean Flows for One-step Generative Modeling (opens in a new window)
- Mean Flows for One-step Generative Modeling (opens in a new window)
- Equilibrium Reasoners: Learning Attractors Enables Scalable Reasoning (opens in a new window)
- About me - Zhengyang Geng (耿正阳) (opens in a new window)