Xinyue Ma

PhD Candidate
Advised by Professor Myeongjae Jeon
Graduate School of Artificial Intelligence
OMNIA at POSTECH [CV]

Publications

  • ORBITFLOW: SLO-Aware Long-Context LLM Serving with Fine-Grained KV Cache Reconfiguration
    Xinyue Ma* , Heelim Hong* , Jongseob Lee , Seoyeong Choy , Woo-Yeon Lee , Taegeon Um , Myeongjae Jeon
    *Equal distribution
    Under Review,
  • REP: Resource-Efficient Prompting for Rehearsal-Free Continual Learning
    Sungho Jeon , Xinyue Ma , Kwang In Kim , Myeongjae Jeon
    NeurIPS 2025, [PDF]
  • Cost-effective On-device Continual Learning over Memory Hierarchy with Miro
    Xinyue Ma , Suyeon Jeong , Minjia Zhang , Di Wang , Jonghyun Choi , Myeongjae Jeon
    ACM MobiCom 2023, [PDF] [Slides] [Code]
  • Experiences

    Web Chair @ SIGOPS APSys 2025

    2025.10.12 - 2024.10.13

    Internship @ RiSE Group

    Microsoft Research Redmond
    2024.9 - 2024.12
    Effective KV Cache Reuse for Retreival-Augmented Generation

    Internship @ Intelligent Cloud Edge Group

    Microsoft Research Asia
    2023.11 - 2024.2

    Teaching Assistant @ POSTECH

    POSTECH
    2024 - Present
    - [CSED703O] AI Systems (2025 Fall)

    Teaching Assistant @ UNIST

    UNIST
    2022 - 2024
    - Computer Architecture (2023 Fall) - Parallel Computing (2022 Fall) - Advanced Programming (2022 Spring)

    Research Interests

    Efficient Systems for LLMs

    Continual Learning

    Skills

    Programming Languages - C/C++, Python, SQL, VDHL, Assembly
    Tools - PyTorch, TensorFlow, MapReduce
    Skillset - Socket Programming, Pintos, Verilog (Behavioral and Structural)