Yifei Zhao Ph.D. Student Computer Science University of Central Florida Orlando, FL Email: yifei.zhao@ucf.edu [Google Scholar] [CV]

Biography

I am a Ph.D. student in Computer Science at the University of Central Florida, co-advised by Prof. Mengxin Zheng and Prof. Qian Lou. My research focuses on secure and efficient AI systems, including IP protection for AI models, trusted execution environments for LLMs, and model efficiency.

Education

• University of Central Florida, Orlando, FL
  Ph.D. in Computer Science, Aug 2024 - Present
• University of Washington, Seattle, WA
  M.S. in Electrical and Computer Engineering, Sep 2023 - Jun 2024

Research Publications

(* indicates equal contribution)

• SIF: Semantically In-Distribution Fingerprints for Large Vision-Language Models
  Yifei Zhao, Qian Lou, Mengxin Zheng
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Denver, Colorado, USA, 2026
  
  TL,DR: Developed an ownership verification framework for open-source vision-language models against unauthorized reuse and API-based redeployment.
• AttestLLM: Efficient Attestation Framework for Billion-scale On-device LLMs
  Ruisi Zhang*, Yifei Zhao*, Neusha Javidnia, Mengxin Zheng, Farinaz Koushanfar
  Design Automation Conference (DAC), Long Beach, California, USA, 2026
  
  TL,DR: Designed and implemented a TEE-based framework for model integrity verification of billion-parameter on-device LLMs on ARM edge platforms.

Preprint Publications

• Secure LLM Execution with TEEs and Crypto-protected Accelerators
  Jiaqi Xue*, Yifei Zhao*, Mengxin Zheng, Fan Yao, Yan Solihin, Qian Lou
  arXiv preprint, 2025
  
  TL,DR: Architected a secure LLM inference framework that offloads about 87% of Transformer computation from TEEs, achieving 3.3x-5.1x speedup via crypto-protected accelerators and heterogeneous execution (Intel SGX + CUDA + llama.cpp).

Technologies

• Frameworks and Tools: PyTorch, vLLM, llama.cpp, ggml
• Trusted Execution Environments: ARM TrustZone, Intel SGX, protected KVM
• Model Efficiency: Model quantization, edge-device inference, performance tuning

Experience

• University of Central Florida
  Teaching Assistant - COP3402: System Software (Aug 2025 - Present)
  Led weekly recitations and office hours; mentored 130+ students on systems programming and compiler fundamentals in C.

Honors and Awards

• ORCGS Doctoral Fellowship, University of Central Florida, 2024