I work on the systems behind large language models — how they train, how they run, and how to make them more efficient.
Currently at Xi’an Jiaotong University, focusing on scalable training/inference, scheduling, and system-level optimizations for LLMs.
Before this, I studied Information Security at Shandong University, with interests in cryptography, secure protocols, and general system security.
This background shaped how I think about reliability, correctness, and the design of trustworthy systems.
I also spent some time at Xiaomi, working on the Android Telephony framework and Phone/InCallUI modules, where I learned to appreciate large codebases and real-world system behaviors
- LLM training & inference infrastructure
- AI accelerator / AI architecture
- system reliability & performance