Changho Hwang

Ph.D. Candidate @ School of Electrical Engineering, KAIST

Room 820, N1 ITC-Building

291 Daehak-ro, Yuseong-gu, Daejeon 34141, Republic of Korea


Graduate student at KAIST, studying high-performance networked systems and deep learning. Working in Networked & Distributed Computing Systems Lab (NDSL) with my advisor, Prof. KyoungSoo Park, since Dec 2015.

Research Interest

Systems support for deep learning, scalable networked systems, system performance optimization


Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea

Ph.D. student, Electrical Engineering

Advisor: KyoungSoo Park
Feb 2017 —

M.S. student, Electrical Engineering

Advisor: KyoungSoo Park
Feb 2016 — Jan 2017

B.S., Electrical Engineering (major) and Computer Science (minor)

Feb 2012 — Jan 2016

Professional Experience

Microsoft Research Asia, Beijing, China

Research Intern, Network Research Group

Jul 2019 — Sep 2019

Dec 2018 — Feb 2019

Research Projects

ARK Project

Jan 2019 —

GPU thread scheduler for high resource utilization and efficient networking in deep learning applications.

CoDDL Project

May 2017 —

Cloud coordinator and scheduler for distributed deep learning.

Confident Multiple Choice Learning

Aug 2016 — Mar 2017

Efficient ensemble method for DNN classifiers that leverages multiple specializer models.
Open Source

APUNet Project

Jan 2015 — Dec 2015

High-performance packet processor accelerated by an integrated GPU.


Confident Multiple Choice Learning

Kimin Lee, Changho Hwang, KyoungSoo Park, and Jinwoo Shin

In Proceedings of the 34th International Conference on Machine Learning (ICML)

Sydney, Austrailia, August 2017

APUNet: Revitalizing GPU as Packet Processing Accelerator

Younghwan Go, Muhammad Jamshed, YoungGyoun Moon, Changho Hwang, and KyoungSoo Park

In Proceedings of the 14th USENIX Symposium on Networked Systems Design and Implementation (NSDI)

Boston, MA, March 2017

Workshops & Posters

A Case for Two-stage Inference with Knowledge Caching

Geonha Park, Changho Hwang, and KyoungSoo Park

In the 3rd International Workshop on Embedded and Mobile Deep Learning (EMDL@MobiSys)

Seoul, Republic of Korea, June 2019

Efficient Resource Sharing for Distributed Deep Learning

Changho Hwang, Taehyun Kim, Kyuho Son, Jinwoo Shin, and KyoungSoo Park

In the 13th USENIX Symposium on Operating Systems Design and Implementation (OSDI) — poster

Carlsbad, CA, October 2018


  • C/C++
  • Unix/GNU Linux
  • Python
  • PyTorch
  • CUDA
  • OpenCL
  • TensorFlow
  • LaTeX

Changho Hwang —