Zihan Zhang | 章子涵

I'm a third-year undergraduate in the ACM Honors Class at Zhiyuan College, Shanghai Jiao Tong University (B.Eng. in CS, expected 2027). GPA: 3.91/4.3 (Rank 8/27).

My research focuses on efficient machine learning systems, particularly model compression and quantization techniques for large language models (LLMs). I aim to develop low-bit methods that deliver high performance with minimal cost, making advanced AI more accessible and sustainable.

Currently researching at EPCC Lab (SJTU, advised by Prof. Jingwen Leng) and remote collaboration with Z-Lab (UC San Diego).

Check my CV or email me!

Email / CV / GitHub / X

News

[Feb 2026] ParoQuant accepted to ICLR 2026! Remote collaboration with Z-Lab (UC San Diego).
[Jan 2026] M2XFP accepted to ASPLOS '26 (EPCC Lab project).
[Oct 2025] Started remote collaboration with Z-Lab (UC San Diego), contributing to efficient LLM quantization methods (ParoQuant).
[Jun 2025] Started research internship at EPCC Lab, focusing on efficient large language models.
[Feb 2025] Became TA for Data Structures (CS1951, Spring 2025).
[Sep 2024] Became TA for Programming (CS1953, Fall 2024).
[Sep 2023] Admitted to ACM Honors Class, Zhiyuan College, SJTU.

Publications

M2XFP: A Metadata-Augmented Microscaling Data Format for Efficient Low-bit Quantization
Weiming Hu, Zihan Zhang, Haoyan Zhang, ... Jingwen Leng
ASPLOS '26 [arXiv] [code]
ParoQuant: Pairwise Rotation Quantization for Efficient Reasoning LLM Inference
Yesheng Liang, Haisheng Chen, Zihan Zhang, Song Han, Zhijian Liu
ICLR 2026 [arXiv] [project] [code]

Research & Education Experience

	Z-Lab (Remote Collaboration), UC San Diego Oct 2025 – Present Advisor: Prof. Zhijian Liu Contributed to ParoQuant, pairwise rotation quantization for reasoning LLMs (accepted to ICLR 2026).
	Emerging Parallel Computer Center (EPCC Lab), SJTU Jun 2025 – Present Research Intern Advisor: Prof. Jingwen Leng Developed M2XFP for low-bit LLM quantization (accepted to ASPLOS '26).
	Shanghai Jiao Tong University (SJTU) Sep 2023 – Present Undergraduate Student, ACM Honors Class GPA: 3.91/4.3 (Rank: 8/27) Academic Advisor: Prof. Yong Yu

Personal Projects

rCore – RISC-V OS kernel in Rust: Sv39 paging, Buddy+Slab allocator, POSIX-like VFS, process scheduling, MMU support.
Mx Compiler & JITC – Full compiler for C-like Mx* language to RISC-V LLVM IR (OOP, f-strings); JIT interpreter for LLVM IR.
RISCV-CPU – Out-of-order RISC-V CPU in Verilog (Tomasulo), FPGA-deployed on Basys3 with cache, UART, exceptions.
Raft Consensus – MIT 6.824 implementation: leader election, log replication, fault tolerance (private repo per policy).

Awards and Honors

2024–2025 & 2023–2024: Zhiyuan Honorary Scholarship

This homepage is adapted from a previous version of Xingyang Li's website, which itself was inspired by the designs of Haoran Geng and Jon Barron. And the background animation is from canvas-nest.js by hustcc
Last updated: Oct 5, 2025
© 2025 Zihan Zhang