Zedong Peng

zedongpeng1@gmail.com zedongpeng.com

PDF

Education

M.Eng., Electronic and Information Engineering, Shanghai Jiao Tong University [transcript]

School of Computer Science. Advised by Prof. Jieru Zhao.

Shanghai, China 2024 – 2027 (expected)

B.Eng., Information Engineering, Shanghai Jiao Tong University [transcript]

School of Electronic Information and Electrical Engineering. Outstanding Graduate.

Shanghai, China 2020 – 2024

Experience

Research Intern, National Center of Technology Innovation for EDA

Advised by Prof. Qiang Xu and Prof. Jieru Zhao.

  • The HLS part of LLM4EDA: LLM post-training and data construction for a domain-specific language.
Nanjing, China July 2024 – 2025

Machine Learning Intern, Intel

Advised by Yuan Zhou.

  • Spark SQL native engine Gluten: optimizing the CI pipeline.
  • LLM framework optimization: Intel device support for vLLM.
  • LLM applications: text to SQL.
Shanghai, China Nov 2023 – July 2024

Winter School, Data Science Institute, Imperial College London

  • Applied vector representations to analyze language features in COVID-19 text.
London, UK Jan 2024 – Feb 2024

Research Intern, Shanghai Jiao Tong University

Advised by Prof. Liyao Xiang.

  • Security-focused machine learning research.
Shanghai, China July 2022 – July 2024

Publications

2026

  1. MemoryTax: The Bitter Lesson of Agent Memory [paper] [arxiv] [code]

    Zedong Peng

    arXiv preprint arXiv:2610.29823

    2026

  2. DiffHLS: Differential Learning for High-Level Synthesis QoR Prediction with GNNs and LLM Code Embeddings [arxiv] [pdf] [code] [poster]

    Zedong Peng, Zeju Li, Qiang Xu, Jieru Zhao

    IEEE/ACM International Symposium of EDA (ISEDA)

    2026

2025

  1. HIPPO: A Hierarchy-Preserving and Noise-Tolerant Pre-HLS Power Modeling Framework for FPGA [paper] [pdf]

    Zefan Lin, Zedong Peng, Mingzhe Gao, Jieru Zhao, Zhe Lin

    Proceedings of the 44th IEEE/ACM International Conference on Computer-Aided Design (ICCAD)

    2025

  2. ForgeHLS: A Large-Scale, Open-Source Dataset for High-Level Synthesis [arxiv] [pdf] [code] [data]

    Zedong Peng, Zeju Li, Mingzhe Gao, Qiang Xu, Chen Zhang, Jieru Zhao

    arXiv preprint arXiv:2507.03255

    2025

  3. DeepCircuitX: A Comprehensive Repository-Level Dataset for RTL Code Understanding, Generation, and PPA Analysis [paper] [arxiv] [pdf] [data]

    Zeju Li, Changran Xu, Zhengyuan Shi, Zedong Peng, Yi Liu, Yunhao Zhou, Lingfeng Zhou, Chengyu Ma, Jianyuan Zhong, Xi Wang, Jieru Zhao, Zhufei Chu, Xiaoyan Yang, Qiang Xu

    IEEE International Conference on LLM-Aided Design (ICLAD)

    2025

2024

  1. A verifiable and privacy-preserving federated learning training framework [paper] [pdf]

    Haohua Duan, Zedong Peng, Liyao Xiang, Yuncong Hu, Bo Li

    IEEE Transactions on Dependable and Secure Computing

    2024

Projects

llama.cpp FPGA backend [blog]

A llama.cpp backend that runs LLM inference on an AMD Alveo U280 FPGA card; twelve models from four families run on it. The code is private for now.

In progress

Robotic Dog Development

Advised by Prof. Weiming Zhao and Xiaomi Technology.

  • Implemented algorithms that improved locomotion control and user interaction.
Shanghai, China Nov 2023 – Nov 2024

Honors and Awards

First-class Postgraduate Academic Scholarship, Shanghai Jiao Tong University

2025

Outstanding Graduate, Shanghai Jiao Tong University

2024

2nd Prize, 13th National University Mathematics Competition

2021

1st Prize, SJTU “Yunhan Cup” RobotMaster Intramural Competition

2021

Liu Yongling Foundation Scholarship

2021

Advanced Individual in Social Practice, Shanghai Jiao Tong University

2021

Skills

Programming: C++, Python, SQL, Verilog, Shell, Docker, MATLAB

Languages: Chinese (native), English (CET-6, 481), German (A1)

Updated October 2026