Huanyi Xie

KAUST, Saudi Arabia

huanyi.xie@kaust.edu.sa

I am an MS/PhD student in Computer Science at King Abdullah University of Science and Technology (KAUST), advised by Prof. Di Wang.

My current work focuses on efficient training and inference of large language models, especially LLM serving, memory-efficient fine-tuning, and data attribution.

Previously, I received my M.Sc. in Data Science from KTH Royal Institute of Technology and my B.Eng. in Artificial Intelligence from Harbin Institute of Technology, Weihai.

I also build small open-source systems for understanding modern LLM infrastructure:

nanoSGLang: a minimal LLM serving engine with continuous batching, paged KV cache, chunked prefill, and streaming inference.
nanoMegatron: a minimal distributed training framework covering DDP, ZeRO, tensor parallelism, pipeline parallelism, and expert parallelism.

You can find more details in my CV or on GitHub.

selected publications

Efficient Text-Attributed Graph Learning through Selective Annotation and Graph Alignment

Huanyi Xie, Lu Yu, Tianhao Huang, and 5 more authors

Transactions on Machine Learning Research (TMLR), 2026
High-Throughput and Memory-Efficient Zeroth-Order Fine-tuning LLMs with Distributed Parallel Computing

Liangyu Wang, Huanyi Xie, and Di Wang

In Annual Meeting of the Association for Computational Linguistics (ACL), 2026
Predicting LLM Output Length via Entropy-Guided Representations

Huanyi Xie, Yubin Chen, Liangyu Wang, and 2 more authors

In International Conference on Learning Representations (ICLR), 2026