Leon Lu

In 2013, Leon Lu started their work experience as a High Performance Heterogeneous Computing professional at OmniVision Technologies, Inc. Leon was responsible for video and image processing algorithms, ARM assembly, NEON, OpenCL, multi-threading, Movidius vector operations, and camera ISP Linux driver development.

In 2014, they moved on to work at Meitu, where they worked as a Linux Kernel and Driver Developer. Leon was in charge of Memory Controller, DDR, Thermal, I2C, SPI Linux driver development and maintenance, ARM-based virtualization, eMCP chip verification, system power and performance optimization, and troubleshooting Linux kernel panic and system hang issues.

From 2016 to 2021, Leon Lu worked at Gao Tong as an NPU Compiler and Runtime Software Designer and Developer. Leon was responsible for designing and developing AI compiler and NPU runtime for Qualcomm Snapdragon NPUs. Leon also self-designed a C++ based AI compiler and used TVM for Hexagon ADSP code generation optimizations.

In 2021, Leon Lu joined NIO, where they worked on a new heterogeneous computing compiler design and development project. Leon acted as the owner for the overall design, development planning, and scheduling of a new LLVM-based compiler for GPGPU-like AI cores in ADAS chips.

Finally, in 2022, Leon Lu joined HeiZhiMo Smart Technology Co., Ltd. as an AI Compiler Manager. Leon is currently working on designing and implementing a high-performance heterogeneous computing AI compiler for ADAS chips, using open-source AI compiler frameworks. Leon is also the founder of the AI compiler technical team in China.

Leon Lu attended the University of Electronic Science and Technology of China from 2010 to 2013, where they obtained a bachelor's degree in Communication and Information Engineering.