Chen Zhang - Associate Research Fellow

Chen Zhang

Associate Research Fellow

Beijing, China

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Chen Zhang is an Associate Research Fellow at Tsinghua University with nine years of research and engineering experience spanning systems, machine learning, and computational neuroscience. He holds a PhD in computational neuroscience and an MS in computer science from USC, and has research stints at Stanford and USC that bridge academic rigor with practical system building. Chen contributes to high-profile open-source projects, notably implementing low-level IOMMU/BTI components in a Rust reimplementation of the Zircon microkernel and adding multi-modal Llama 3.2 support and performance optimizations to the VLLM inference engine. His work sits at the intersection of kernel/device-driver internals and efficient LLM inference, bringing rare expertise in both hardware-adjacent systems and modern ML model engineering. Colleagues describe him as a meticulous implementer who surfaces subtle kernel and kernel-ML interactions that materially improve performance and reliability.

9 years of coding experience

9 years of employment as a software developer

Doctor of Philosophy (PhD), Neuroscience (Computational track), Doctor of Philosophy (PhD), Neuroscience (Computational track) at University of Southern California

Bachelor's degree, BME, Bachelor's degree, BME at Tsinghua University

English, Chinese

Github Skills (19)

pytorch10

python10

iommu10

rc10

inference10

microkernel10

llm10

flash10

system-calls10

mlops10

rcon10

self-attention10

cuda10

transformer10

multihead-attention10

Programming languages (9)

TypeScriptC++ShellRustMakefileTeXHTMLPython

Github contributions (5)

vllm-project/vllm

Sep 2024 - Apr 2025

A high-throughput and memory-efficient inference and serving engine for LLMs

Role in this project:

ML Engineer

Contributions:90 reviews, 34 PRs, 1 push in 6 months

Contributions summary:Chen primarily contributed to the Mllama model within the VLLM project, adding support for multi-modal Llama 3.2, including handling image inputs and related processing. They fixed bugs in the benchmark throughput and encoder-decoder functionality, ensuring prefill throughput testing with specific configurations. Their work also involved kernel improvements, such as supporting sliding windows in the flash attention backend, and enhancements to prefix caching for efficient model execution.

amdcudadeepseekgpthpu

rcore-os/zCore

Apr 2020 - Jun 2020

Reimplement Zircon microkernel in Rust.

Role in this project:

Back-end Developer

Contributions:1 review, 53 commits, 7 PRs in 1 month

Contributions summary:Chen implemented core functionalities of the Zircon microkernel, focusing on IOMMU and BTI (Bus Transaction Initiator) components. Their contributions included the creation and management of IOMMU and BTI objects, incorporating related system calls. They also worked on device-driver-related functions. Additionally, the user made modifications in the interrupt system.

microkernelrustzirconembeddedreimplement

Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.

Request Free Trial