Postdoctoral Research Fellow

Zexin Li

Zexin Li

I am a postdoctoral research fellow at Nanyang Technological University, where I work with Dr. Arvind Easwaran. I received a Ph.D. degree from the University of California, Riverside (UCR) under the advice of Dr. Cong Liu in 2026. I also received a master’s degree from UCR in 2024. I received a bachelor’s degree from the Southern University of Science and Technology (SUSTech) under the advice of Dr. Yuqun Zhang in July 2020. My research interests lie in interdisciplinary fields of real-time embedded systems and on-device machine learning.

I am actively looking for cooperation in the following topics:

  1. deploying interactive AI models on resource constrained edge devices
  2. co-scheduling inference and model updates within resource constraints and response deadlines
  3. system failure detection, safe fallback and recovery
  4. consistency verification between pre- and post-deployed model pipelines

Feel free to contact me if we share common research interests.

News

Paper acceptances (Feb 2023–Aug 2026)
Archive News
  • One paper is accepted by TMLR.
  • Honored to receive the Dissertation Completion Fellowship Award (DCFA), UCR.
  • Our RTSS'25 paper received Outstanding Paper Award.
  • One paper is accepted by RTSS'25.
  • Two papers are accepted by RTSS'24.
  • Three papers are accepted by RTSS'23.
  • Honored to receive the Dean’s Distinguished Fellowship, UCR.
  • Join UCRiverside and Dr. Cong Liu’s Autonomous Embedded Systems Lab in 2022 Fall.
  • Join UTDallas and Dr. Cong Liu’s Real-Time System Lab in 2022 Spring.
  • Begin the research intern in Tecent Youtu Lab, Shanghai, China; Admitted into Tencent Rhino Bird Elite Talent Training Program.
  • One paper is accepted by JSA, great thanks to Dr. Yuqun Zhang.
  • Begin the research intern in Kuaishou, Shenzhen, China.

Selected Publications

Full publications

ProCerDCL: Query-Efficient Black-Box Probabilistic Verification of Deep Learning Compiler Optimization
arXiv preprint
Real-Time Continual Learning on Embedded GPUs
Artifacts Evaluated RTSS'26
MixTraining: A Better Trade-Off Between Compute and Performance
TMLR'26
LeMix: Unified Scheduling for LLM Training and Inference on Multi-GPU Systems
Outstanding Paper Award RTSS'25
R³: On-device Real-Time Deep Reinforcement Learning for Autonomous Robotics
RTSS'23
RED: A Systematic Real-Time Scheduling Approach for Robotic Environmental Dynamics
RTSS'23

Leadership and Service