One paper accepted by IEEE TC.
Shengyu Fan
Assistant Professor
School of Artificial Intelligence
Shandong University
Jinan & Beijing, China
Biography
I am an Assistant Professor in the School of Artificial Intelligence at Shandong University. I received my PhD from the Institute of Information Engineering, Chinese Academy of Sciences.
My research focuses on efficient and trustworthy next-generation AI systems, with an emphasis on computer architecture and systems software. My work spans fully homomorphic encryption, GPU systems, and specialized accelerators. I value dialogue with industry as a way to identify research questions grounded in real applications and systems.
Prospective students
I welcome PhD, master’s, and undergraduate students interested in computer architecture, fully homomorphic encryption, and AI systems.
Please email me with a brief introduction and your research interests: damionfan@163.com
News
Serving on the USENIX Security '27 Program Committee.
Serving on the HPCA 2027 Program Committee.
Two papers accepted by IEEE TCAD. Thanks to all collaborators.
Professional Service
Program Committees
- 2027: USENIX Security, HPCA
- 2026: MICRO
Journal Reviewing
IEEE TCAD, IEEE TC, ACM TACO
Artifact Evaluation Committees
- 2025: EuroSys, SOSP, MICRO, HPCA
Research
My research focuses on efficient and trustworthy next-generation AI systems. I study inference execution, privacy-preserving computation, and computing architectures, bringing together algorithms, systems software, and hardware design.
Efficient AI Systems
I study computation, memory access, and scheduling in AI inference. My interests include GPU operator execution, data reuse, and resource management to reduce inference latency and resource costs.
Trustworthy AI Computing
I study privacy-preserving computation and secure execution on sensitive data, with a focus on fully homomorphic encryption and its system support. The goal is to improve efficiency under explicit security constraints.
Computer Architectures for AI
I study GPUs, specialized accelerators, and heterogeneous architectures for AI and privacy-preserving computation. This work connects compute units, memory hierarchies, and dataflows to improve hardware efficiency and support different workloads.
Selected Publications
CACO: Chiplet-based FHE Acceleration with Inter-Chiplet Communication Optimizations
OmniCifro: A Heterogeneous Acceleration System with PIM for Encrypted Computation
MNEMOS: A GPU-based TFHE Acceleration Framework with Memory Access Optimization
HyperDrive: Hierarchical Exploitation of Memory Efficiency for GPU-Based FHE Acceleration
Falcon: Algorithm-Hardware Co-Design for Efficient Fully Homomorphic Encryption Accelerator
Emaildamionfan@163.com
Based inJinan & Beijing, China