I’m a deep learning systems engineer–scientist interested in making modern AI models actually work in the messy, heterogeneous deep learning stacks they run on in the real world—across models, systems, and hardware.
My research focuses on heterogeneous deep learning: designing execution algorithms and runtimes that jointly co-optimize model architectures, system software, and accelerator characteristics while remaining compatible with existing deep learning ecosystems.
As an engineer, I build kernels, runtimes, and tools that people can actually use. As a scientist, I try to extract the general principles behind these systems: how models, compilers, and hardware should co-evolve in heterogeneous environments. In the long run, I want to help shape a stack where model, system, and hardware are not separate worlds but a single design space—from low-level optimizations and scheduling policies up to hardware-aware yet portable model architectures.
Publications - [link]
Carrer -[링크]
Extra Curricular Study - [링크]
Projects -[링크]
blog1: Woongjoon AI blog2 : 티스토리 블로그 blog3(english) : Woongjoon_AI2