Zongpu Zhang | SJTU

He is a fifth-year Ph.D. student advised by Prof. Haibing Guan at the Department of Computer Science, Shanghai Jiao Tong University. He finished the Bachelor’s degree at the Department of Computer Science, Shanghai Jiao Tong University in 2019. Now he is a visiting scholar advised by Prof. Xuehai Qian at the Department of Computer Science, Purdue University.

He’s on the job market for a post-doc position in 2024.

Research Interests

• I/O Virtualization Systems with software-hardware co-design.
  By collaborating with Intel, we built a highly scalable software-hardware co-designed I/O virtualization system that achieves at most 3x more virtual devices supporting the same performance as the traditional SR-IOV solution (HD-IOV, to be published in EuroSys’24). We also built a para-virtualization stack for crypto accelerators that achieves near-native I/O performance leveraging techniques such as DPDK (vCrypto, to be published in INFOCOM’24).
• In-network Offloading Systems with programmable devices such as DPUs and P4 switches.
  We offloaded crypto operations to dedicated accelerators and built a cloud Key Management System (KMS) that achieves hardware-level key protection and up to 17x higher asymmetric crypto algorithm throughput compared with software solutions (QKPT, in IEEE TDSC’21). We also offloaded virtio back-ends to the hardware device and designed a live migration logic to facilitate migratable pass-through devices (Un-IOV, in IEEE TC’24). We are actively developing DPU offloaded far memory systems and P4 offloaded transaction processing systems.
• Machine Learning Systems for Large Language Models (LLM).
  We are exploring system-level opportunities to optimize LLM finetuning and serving efficiency, using techniques such as pruning, quantization, parallelism, offloading, and memory management.

Professional Experience

• Reviewer for the 2024 International Journal of Computer and Telecommunications Networking

Industrial Experience

• Software Development Intern (NPG), Intel Asia-Pacific R&D Ltd.
  March, 2020 - March, 2023
• Research Intern (ADAS team), Sensetime Inc.
  Summer 2019
• Deep Learning Intern (DCG), Intel Asia-Pacific R&D Ltd.
  Summer 2018

Teaching Experience

• 2020,21,22, Teaching assistant, EI313 Science and Technology Innovation, Shanghai Jiao Tong University

Honors and Award

• 2019, Zhiyuan Honored Ph.D. Program, Shanghai Jiao Tong University
• 2019, Excellent Bachelor’s Thesis Award (Top 1%), Shanghai Jiao Tong University
• 2019, Zhang Xu Scholarship, SEIEE, Shanghai Jiao Tong University
• 2019, Yang Yuanqing Scholarship, SEIEE, Shanghai Jiao Tong University
• 2018, SenseTime Scholarship, Sensetime Inc.
• 2016,17,18, Academic Excellence Scholarship (three times), Shanghai Jiao Tong University