About me

Table of Contents

About me

Thank you for your interest.

I am a Deep Learning system architect at NVIDIA, and my current focus is high-performance AI Compiler(on GPU).

Before this, I was a senior engineer in Baidu, working as an architect on PaddlePaddle (one of the most popular open-sourced deep learning frameworks in China market).

I was the creator & primary author & tech lead of the following projects in PaddlePaddle ecosystem (before 2022-6)

• Paddle-Inference - the server-class DL inference engine,
• Paddle-Lite - A on-device inference engine with high-performance, tiny deployment size,
• Paddle-CINN - A DL compiler for automatically generating high-performance kernels & programs for AI models,
• Paddle/infrt - A unified architecture for both server-class and mobile devices, powered by MLIR, highly modular design,
• Paddle-VisualDL - A visualization tool for AI model,
• Paddle/continuous_evaluation - A system that continuously traces the indicators modeling the performance of AI models.