Search

Tianle Zhong

LLMPROF: Identifying Performance Bottlenecks in LLM Serving Systems with Top-Down Profiling
Paper Accepted at SC 2026
Proton: Towards Multi-level, Adaptive Profiling for Triton

We build compiler, architecture, and toolkit foundations for faster, more understandable intelligent systems.

Computer Science · George Mason University · Fairfax, Virginia