I’m a software architect. I’ve work extensively with deep learning, from hardware architecture to applications. I will receive a Ph.D. degree in computer science and engineering from Seoul National University, Seoul, Korea, in Feb 2021. I am interested in the deep learning and the HW/SW co-design. I am working on an neural processing unit (NPU) compiler.
Improve our NPU architecture to support several non-convolutional deep learning operations. Build and optimize the prototypes of multiple NPU systems.
Propose a new NPU architecture (MIDAP), implement a simulator, and conduct research on the compiler and the micro architecture.
Methodology for scheduling deep learning applications with heterogeneous hardware accelerators
Deep learning system for speaker recognition
Optimization methodology of object detection deep learning application for autonomous vehicles
Optimization of deep learning for the embedded system
Implement the previous version of embedded accelerator simulator and optimize memory architecture on the accelerator.
A study on fault injection method at the Linux kernel level using KGDB
Jan 2014 - Feb 2021, Seoul in South Korea
Mar 2016 - Feb 2021
Mar 2014 - Feb 2016
Jan 2014 - Feb 2014