Sangjin Choi

About

Hello, I am a Ph.D. student at KAIST, specializing in Machine Learning Systems. My research primarily centers on enhancing the performance of distributed training and inference in ML systems. I bring unique insights into system optimization by harnessing the capabilities of underlying hardware. My passion lies in advancing sustainable and efficient development in ML systems.

Education

• Ph.D. Student in School of Computing, KAIST, current
• M.S. in School of Computing, KAIST, 2021
• B.S. in School of Computing, KAIST, 2019

Publications

• [ICCAD 2023] Jaehoon Heo, Yongwon Shin, Sangjin Choi, Sungwoong Yune, Jung-Hoon Kim, Hyojin Sung, Youngjin Kwon, Joo-Young Kim, PRIMO: A Full-Stack Processing-in-DRAM Emulation Framework for Machine Learning Workloads (Acceptance rate: 22.9%) [ Paper ]
• [USENIX ATC 2023] Sangjin Choi, Inhoe Koo, Jeongseob Ahn, Myeongjae Jeon, Youngjin Kwon, EnvPipe: Performance-preserving DNN Training Framework for Saving Energy (Acceptance rate: 18.4%) [ Paper | Code | Slides | Talk ]
• [USENIX ATC 2022] Sangjin Choi*, Taeksoo Kim*, Jinwoo Jeong, Rachata Ausavarungniurn, Myeongjae Jeon, Youngjin Kwon, Jeongseob Ahn, Memory Harvesting in Multi-GPU Systems with Hierarchical Unified Virtual Memory (*Co-first author, Acceptance rate: 16.2%) [ Paper | Code | Slides | Talk ]

Other Readings

• THE FINE-TUNING DESIGN ARGUMENT by Robin Collins