🤖 Brief Introduction

Hello, I am JamesNULLiu (Yanchen Liu, 刘彦辰).

I am a first-year MSCS student at University of Southern California, now working as a research intern in INK Lab @ USC, advised by Prof. Xiang Ren and Postdoc. Siyuan Wang.

My current research focuses on LLM post-training and reasoning, particularly in areas such as SFT, RLHF, and test-time reasoning. I am also interested in improving the efficiency of LLM training and inference, as well as the development of LLM-based agents.

More broadly, my goal is to improve the efficiency and capability of LLMs at both the algorithmic and systems levels. At the algorithmic level, I aim to enhance model reasoning through improved training data, reward modeling, and sampling strategies during post-training. At the systems level, I am interested in optimizing training and inference frameworks to better utilize heterogeneous hardware and improve the scalability and efficiency of LLM deployment.

If you have any inquiries or are interested in collaboration, please feel free to contact me via email at jamesnulliu@gmail.com.

🧑‍🎓 Education History

  • 2025.09 - 2027.06: Master of Computer Science, University of Southern California, Los Angeles, California, USA.
  • 2021.09 - 2025.06: Bachelor of Computer Science, Shanghai University, Shanghai, China.

💻 Professional Experience

  • 2025.07 - Present: Graduate Research Intern, INK Lab, University of Southern California.
  • 2024.07 - 2025.06: MLE Intern, Shanghai AI Laboratory.
  • 2023.03 - 2025.04: Undergraduate Research Assistant, SHUCS Lab, Shanghai University.
  • 2023.03 - 2025.04: Team Leader, Shangai University Super Computing Team.
  • 2023.06 - 2024.07: Undergraduate Research Intern, Shanghai University and East China Air Traffic Control Bureau.

🎉 Honors and Awards

  • [2024.04] First Prize and Group Competition Award in 2024 ASC Student Supercomputer Challenge Global Final.
  • [2022.06] First-Class Academic Scholarship for outstanding academic performance, Shanghai University.

📰 Publications

2026 ————

  1. [ICLR] S. Wang, Y. Liu, X. Ren. “Segment-Level Attribution for Selective Learning of Long Reasoning Traces”.

2025 ————

  1. [Arxiv | Code] J. Lv, X. He, Y. Liu, A. Shen, X. Dai$^*$, Y. Li, J. Hao, J. Ding, Y. Hu, S. Yin. “HPCTransCompile: An AI Compiler Generated Dataset for High-Performance CUDA Transpilation and LLM Preliminary Exploration”.
  2. [Arxiv | Code] Q. Liu$^\dagger$, Y. Liu$^\dagger$, R. Li, C. Cao, Y. Li$^*$, X. Li$^*$, P. Wang, R. Feng. “MDHP-Net: Detecting an Emerging Time-exciting Threat in IVN”.
  3. Z. Xu, A. Shen, D. Kong, X. Dai, J. Liu, Y. Liu, L. Wang, S. Wei, Y. Hu and S. Yin*. “LLMEngine: Disaggregated Mapping and Memory Management Co-scheduling for Wafer-scale Chips”.

2024 ————

  1. [IEEE Internet of Things Journal | Code] Q. Liu, X. Li, K. Sun, Y. Li$^*$ and Y. Liu$^*$. “SISSA: Real-Time Monitoring of Hardware Functional Safety and Cybersecurity With In-Vehicle SOME/IP Ethernet Traffic”.
  2. [MDPI Future Internet | Code] X. Li, R. Li, and Y. Liu. “HP-LSTM: Hawkes Process–LSTM-Based Detection of DDoS Attack for In-Vehicle Network”.

🤪 Hobbies

  • 🧙 Animations, Comics and Games
  • 🎸 Electric Guitar
  • 🎼 Jazz, Fusion, Metal, Core, Djent