Zihao Ye (叶子豪) - University of Washington

About Me

This is Zihao Ye, a Ph.D. candidate advised by Luis Ceze. I work closely with Tianqi Chen on Machine Learning Compilers. I'm a research intern at NVIDIA, working with Vinod Grover.

I am visiting CMU Catalyst, hosted by Tianqi Chen.

Prior to my PhD, I spent two years at AWS where Minjie Wang and Zheng Zhang introduced me to the Machine Learning Systems. I obtained my bachelor's degree from ACM Honors Class Shanghai Jiao Tong University.

The easiest way to find me is to create an issue or Pull Request and ping me on Github. I don't reply to emails about job opportunities.

Research

My current research focuses on Machine Learning Compilers and Sparse Computation. I'm passionate about building practical systems that have real-world impact and I enjoy dealing with engineering challenges.

I'm focusing on developing FlashInfer and its compiler and runtime, feel free to drop me an email if are interested, and I'm open to collaborations.

Active Projects

FlashInfer: Kernel Generator Engine for LLM Inference Serving

I'm excited to be part of MLC community and collaborate with a strong team on the following projects in TVM Unity:

MLC-LLM: Universal Deployment of Large Language Models
Relax: Composable Abstractions for End-to-End Dynamic Machine Learning
TensorIR: Tensor-Level Abstractions for Deep Learning Operators

Earlier Projects

Punica: Serving multiple LoRA fineuned LLM as one
SparseTIR: Compiler Abstraction for Sparse Operators in Deep Learning, built upon Apache TVM
DGL: Efficient and Scalable Deep Learning on Graphs

Misc

Heaven Sent is my favorite Doctor Who episode and it helped me get through hard times over the years.