Hi! I am Yaojie Zhang, a senior undergraduate majoring in Software Engineering at the University of Electronic Science and Technology of China (UESTC). In Fall 2027, I will join the School of Computer Science at Peking University as an M.S. student, advised by Prof. Xupeng Miao. Previously, I was a research intern at the EPIC Lab, Shanghai Jiao Tong University, advised by Prof. Linfeng Zhang.
My research focuses on efficient inference and serving for large language models, including both autoregressive language models and diffusion language models. I am particularly interested in speculative decoding, feature caching, and cache optimization. My goal is to develop practical algorithms and systems that accelerate language model generation, reduce memory usage, and preserve output quality.
🔥 News
- 2026.09: 🎉 Our paper “Domino” was accepted to EMNLP 2026.
- 2026.09: 🎉 We released the paper “DFlow: Enabling Verifier Information Flow in Block Diffusion Speculative Decoding”.
- 2026.07: 🎉 Our paper “Mask Tokens as Prophet” was published in Findings of ACL 2026.
- 2026.05: 🎉 Our paper “dLLM-Cache” was accepted to ICML 2026.
- 2026.05: 🎉 We released the papers “FlexDraft” and “Domino”.
- 2026.02: 🎉 Our paper “SlowFast Sampling” was accepted to ICLR 2026.
- 2026.01: 🎉 We released Innovator-VL, a fully open-source multimodal large language model for scientific discovery.
- 2025.10: 🎉 We released the paper “Mask Tokens as Prophet: Fine-Grained Cache Eviction for Efficient dLLM Inference”.
- 2025.07: 🎉 We released the paper “Accelerating Diffusion Large Language Models with SlowFast Sampling: The Three Golden Principles”.
- 2025.05: 🎉 We released the paper “dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching”.
- 2025.03: 🤗 Started a research internship at EPIC Lab, Shanghai Jiao Tong University, focusing on efficient inference methods.
📝 Publications (* denotes equal contribution.)
DFlow: Enabling Verifier Information Flow in Block Diffusion Speculative Decoding
Yaojie Zhang, Linfeng Zhang, Bin Cui, Xupeng Miao.
Domino: Decoupling Causal Modeling from Autoregressive Drafting in Speculative Decoding
Jianuo Huang *, Yaojie Zhang *, Qituan Zhang, Hao Lin, Hanlin Xu, Linfeng Zhang.
FlexDraft: Flexible Speculative Decoding via Attention Tuning and Bonus-Guided Calibration
Yaojie Zhang *, Jianuo Huang *, Junlong Ke, Yuhang Han, Yongji Long, Tianchen Zhao, Biqing Qi, Linfeng Zhang.
Mask Tokens as Prophet: Fine-Grained Cache Eviction for Efficient dLLM Inference
Jianuo Huang *, Yaojie Zhang *, Yicun Yang, Benhao Huang, Linfeng Zhang.
Accelerating Diffusion Large Language Models with SlowFast Sampling: The Three Golden Principles
Qingyan Wei, Yaojie Zhang, Zhiyuan Liu, Puyu Zeng, Yuxuan Wang, Biqing Qi, Dongrui Liu, Linfeng Zhang.
dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching
Zhiyuan Liu *, Yicun Yang *, Yaojie Zhang, Junjie Chen, Chang Zou, Qingyan Wei, Shaobo Wang, Yichen Zhu, Linfeng Zhang.
🛠 Open-source Projects
Innovator-VL: A Multimodal Large Language Model for Scientific Discovery
Zichen Wen *, Boxue Yang *, Shuang Chen, Yaojie Zhang, et al.
[Paper] [Project Page]
🎖 Honors and Awards
- 2026 CCF Elite Collegiate Award (CCF优秀大学生; 103 awardees nationwide)
- 2025 Sichuan Provincial College Student Comprehensive Quality A Level Certificate (四川省大学生“综合素质A级证书”)
- 2025 National Scholarship (国家奖学金; Top 0.4% nationwide)
- 2025 Tencent Scholarship (腾讯企业奖学金)
- 2024 National Scholarship (国家奖学金; Top 0.4% nationwide)
- 2024 Virtuos Corporate Scholarship (维塔士企业奖学金)
📖 Education
- 2027.09 - 2030.06, Peking University, M.S. in Computer Science (Advised by Prof. Xupeng Miao)
- 2023.08 - 2027.06, University of Electronic Science and Technology of China, B.E. in Software Engineering
💻 Internships
- 2025.03 - 2026.07, Research Intern, EPIC Lab, Shanghai Jiao Tong University (Advised by Prof. Linfeng Zhang).