Jiamu Morris Zhang

Duncan Hall, 2090

6100 Main St

Houston, Texas 77005

Seeking a Summer 2027 research internship — efficient LLM inference, agentic systems, and evaluation. mz81@rice.edu

About me

I’m Jiamu ‘Morris’ Zhang (张佳慕), a third-year Ph.D. student in Computer Science at Rice University, in the Chili Lab with Professor Hanjie Chen.

I work on making LLMs and agentic systems cheaper to run, and on measuring whether the savings are real. Most of it is system–method co-design: compression, sparsity and test-time compute control, paired with the systems that actually serve them. See my publications and code, or my Google Scholar.

Before Rice I received my B.S. from Case Western Reserve University, advised by Shuai Xu and Vipin Chaudhary, and was fortunate to be mentored by Dr. Xia “Ben” Hu. Feel free to reach out.

Software

AnyJev turns any LLM into a typed decision model with calibrated probabilities, no fine-tuning required. I lead it at Nokia Applied Research; it installs from PyPI and serves through vLLM. On the community Jev Decision Index (v0.4.1), its fine-tuned variant AnyJev Tacitus is joint 3rd of 219, ahead of the Jev reference. All code →

news

Oct 27, 2026	Our EMNLP 2026 paper Rethinking the Evaluation of Efficiency Methods for Multi-Agent Systems received an Outstanding Paper Award.
Oct 21, 2026	AnyJev Tacitus, the fine-tuned variant, reached joint 3rd of 219 on the community Jev Decision Index (v0.4.1), ahead of the Jev reference. See the board →
Oct 12, 2026	AnyJev is the highest-placed training-free system on the community Jev Decision Index (v0.3.2): joint 4th of 148, ahead of the Jev reference itself, with no fine-tuning and no weight changes. See the board →
Sep 25, 2026	Rethinking the Evaluation of Efficiency Methods for Multi-Agent Systems was selected for an oral presentation at EMNLP 2026.
Sep 21, 2026	Released AnyJev with Nokia Applied Research — turn any LLM into a typed decision model with calibrated probabilities, no fine-tuning. It passed 500 GitHub stars in its first three days.

selected publications

EMNLP 2026 Oral

Rethinking the Evaluation of Efficiency Methods for Multi-Agent Systems

Outstanding Paper Award

Jiamu Zhang^*, Lingxi Zhang^*, Pengjun Lu^*, Qiyue Zhang, and 5 more authors

In EMNLP, Oral , 2026

arXiv Bib

@inproceedings{zhang2026maseval,
  title = {Rethinking the Evaluation of Efficiency Methods for Multi-Agent Systems},
  author = {Zhang, Jiamu and Zhang, Lingxi and Lu, Pengjun and Zhang, Qiyue and Chuang, Yu-Neng and Li, Zhengchen and Xu, Shuai and Chaudhary, Vipin and Chen, Hanjie},
  booktitle = {EMNLP},
  year = {2026},
}

Preprint

BellowsKV: One Latent for Many Layers, Expanded on Demand

Jiamu Zhang and Hanjie Chen

arXiv preprint arXiv:2610.35023, 2026

arXiv Bib

@article{zhang2026bellowskv,
  title = {BellowsKV: One Latent for Many Layers, Expanded on Demand},
  author = {Zhang, Jiamu and Chen, Hanjie},
  journal = {arXiv preprint arXiv:2610.35023},
  year = {2026},
}

CVPR

Flexible Group Count Enables Hassle-Free Structured Pruning

Jiamu Zhang^*, Shaochen Zhong^*, Andrew Ye, Zirui Liu, and 8 more authors

In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025

Bib PDF Code

@inproceedings{zhang2025flexgroup,
  title = {Flexible Group Count Enables Hassle-Free Structured Pruning},
  author = {Zhang, Jiamu and Zhong, Shaochen and Ye, Andrew and Liu, Zirui and Zhao, Sebastian and Zhou, Kaixiong and Li, Li and Choi, Soo-Hyun and Chen, Rui and Hu, Xia and Xu, Shuai and Chaudhary, Vipin},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2025},
}

TMLR

Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models

Yang Sui, Yu-Neng Chuang, Guanchu Wang, Jiamu Zhang, and 7 more authors

Transactions on Machine Learning Research (TMLR), 2025

arXiv Bib Code

@article{sui2025overthinking,
  title = {Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models},
  author = {Sui, Yang and Chuang, Yu-Neng and Wang, Guanchu and Zhang, Jiamu and Zhang, Tianyi and Yuan, Jiayi and Liu, Hongyi and Wen, Andrew and Zhong, Shaochen and Chen, Hanjie and Hu, Xia},
  journal = {Transactions on Machine Learning Research (TMLR)},
  year = {2025},
}