About Me

Hi there! I am Zhongheng Liu, an undergraduate student in Electronic Information and Technology at Tsinghua University (expected graduation in June 2027), selected for the “AI + System” Talent Program.

I recently completed an exchange semester at the School of Electrical and Electronic Engineering (EEE), Nanyang Technological University (NTU) in Singapore. I am currently working on memory compression for Large Video Models with MMLab@NTU under the supervision of Prof. Ziwei Liu. In Summer 2026, I will join HKUST for a summer research program.

My research broadly centers on efficient and adaptive AI systems, especially LLM agents, multimodal AI, graph learning, reinforcement learning, and combinatorial optimization. Previously, I worked with Prof. Quanming Yao at Tsinghua on LLM-based reflective evolution for dynamic network routing, contributing to NeRM-Net and LERNA-Flow.

You can find my full CV here: Download CV.

🔥News

• [Jun. 2026] I will join HKUST for a summer research program.
• [May 2026] Completed my exchange program at NTU EEE.
• [Feb. 2026] Joined MMLab@NTU to work on memory compression for Large Video Models.
• [Oct. 2025] Won 3rd Prize at the 1st AI Agent Innovation Competition at Tsinghua University.
• [Sep. 2025] Our paper “NeRM-Net: Reflective Evolution of Routing Strategies for Dynamic Communication Networks” was accepted at the ACM CoNEXT 2025 Workshop.

📖Education

• Tsinghua University, B.S. in Electronic Information and Technology, Aug. 2023 - Jun. 2027 (Expected)
  • GPA: 3.8/4.0; selected member of the “AI + System” Talent Program.
• Nanyang Technological University (NTU), Exchange Student, School of Electrical and Electronic Engineering, Aug. 2025 - May 2026
  • Completed a fully English-taught exchange program with coursework in image processing, computer communications, artificial intelligence, operating systems, pattern recognition, and deep learning.

Research Interests

• Efficient multimodal AI systems, including memory compression for Video LLMs and KV-cache optimization.
• LLM agents and algorithm evolution for combinatorial optimization.
• Graph learning and dynamic network routing.
• Reinforcement learning for adaptive decision-making.

🎖Honors & Awards

• 3rd Prize, twice, Ministry of Education Scholarship for Students from Hong Kong, Macao and Overseas Chi- nese, 2025|2024
• Zheng Geru Scholarship for Tech Innovation, Tsinghua University, Oct. 2024.
• 3rd Prize, 1st AI Agent Innovation Competition, Tsinghua University, Oct. 2025.
• Honorable Mention, MCM/ICM Mathematical Contest in Modeling, Feb. 2025.
• Winner Prize (Top 30%), 27th Hardware Design Competition, Tsinghua University, Sept. 2024.
• 1st Prize & Best Sharing Award, Tsinghua “Zhili Cup” Agent Competition, Aug. 2024.
• 1st Prize, 7th Software Design Competition (“Bambu Lab Cup”), Tsinghua University, Apr. 2024.