Hongming Zhang

Research Scientist at FAIR, Meta

I am a Research Scientist at FAIR, Meta. My research interests lie in self-evolving agents and structured reasoning models. Recently, I have been working on the agentic capabilities of the Muse Spark model.

Before joining Meta, I was a senior researcher and research lead at Tencent AI Lab. I was also a research scholar at the University of Pennsylvania, working with Prof. Dan Roth. I received my Ph.D. in Computer Science from HKUST in 2021, advised by Prof. Yangqiu Song, and previously earned my M.Phil. and Bachelor's degrees from HKUST.

Projects

News

July 2025: Joined FAIR, Meta as a Research Scientist.
Jan. 2025: Two papers were accepted to ICLR 2025, including work on MLE agents and code agents.
Sep. 2024: Five papers were accepted to EMNLP 2024 and Findings, spanning theory of mind, RAG, and reasoning.
May 2024: Five papers were accepted to ACL 2024 and Findings, with a strong focus on LLM reasoning and agents.
Dec. 2023: Work on sentence representation received the EMNLP 2023 Outstanding Paper Award.

Selected Recent Publications

For a more complete list, see the publications page, or my Google Scholar page.

Self-Evolving Agents

Cognitive Kernel: An Open-source Agent System towards Generalist Autopilots. Hongming Zhang, Xiaoman Pan, Hongwei Wang, Kaixin Ma, Wenhao Yu, and Dong Yu. NAACL 2025 Demo. Introduces a general-purpose agent system for realizing self-evolving agents in real-world tasks. paper code
R-Zero: Self-Evolving Reasoning LLM from Zero Data. Chengsong Huang, Wenhao Yu, Xiaoyang Wang, Hongming Zhang, Zongxia Li, Ruosen Li, Jiaxin Huang, Haitao Mi, and Dong Yu. ICLR 2026. Explores a novel approach to optimizing a model's task generation capability through self-evolution. paper
WebEvolver: Enhancing Web Agent Self-Improvement with Co-evolving World Model. Tianqing Fang, Hongming Zhang, Zhisong Zhang, Kaixin Ma, Wenhao Yu, Haitao Mi, and Dong Yu. EMNLP 2025. Applies self-evolving algorithms to the agentic domain through a co-evolving world model. paper

Structured Reasoning Models

Parallel-R1: Towards Parallel Thinking via Reinforcement Learning. Tong Zheng, Hongming Zhang, Wenhao Yu, Xiaoyang Wang, He Xing, Runpeng Dai, Rui Liu, Huiwen Bao, Chengsong Huang, Heng Huang, and Dong Yu. ICLR 2026. Introduces reinforcement learning for parallel thinking in structured reasoning models. paper code
Streaming Looking Ahead with Token-level Self-reward. Hongming Zhang, Ruixin Hong, and Dong Yu. Technical Report 2025. Proposes token-level self-reward for efficient look-ahead reasoning in streaming settings.
Scaling Test-Time Compute for Agentic Coding. Joongwon Kim, Wannan Yang, Kelvin Niu, Hongming Zhang, Yun Zhu, Eryk Helenowski, Ruan Silva, Zhengxing Chen, Srinivasan Iyer, Manzil Zaheer, Daniel Fried, Hannaneh Hajishirzi, Sanjeev Arora, Gabriel Synnaeve, Ruslan Salakhutdinov, and Anirudh Goyal. In submission to COLM 2026. Studies how to scale test-time compute for structured reasoning in agentic coding tasks.

Google Scholar / GitHub / LinkedIn / Twitter