NeurIPS 2026
Publications
8 publications · conference papers and preprints
Conference papers
ICML 2026
Efficient Distributed MLLM Training with Cornstarch
ICLR 2026
Mordal: Automated Pretrained Model Selection for Vision Language Models
SOSP 2024
Perseus: Reducing Energy Bloat in Large Model Training
SOSP 2023
Oobleck: Resilient Distributed Training of Large Models Using Pipeline Templates
SOSP 2021Best Paper Award
LineFS: Efficient SmartNIC Offload of a Distributed File System with Pipeline Parallelism
ASPLOS 2019
Heterogeneous Isolated Execution for Commodity GPUs
Preprints
arXiv 2026
Addressing Variable Heterogeneity in Distributed Multimodal Training with Entrain
Experience
Academia
University of Michigan · Graduate Student Research Assistant (GSRA)
ML systems for distributed training
KAIST · Research Assistant
RDMA and SmartNIC
KAIST · Research Assistant
PCIe architecture and hardware-based trusted execution environment for GPUs
Industry
NVIDIA · Research Scientist
AI Systems
Google · Software Engineering Intern
Straggler detection in Kubernetes based distributed training
Tesla · Autopilot Software Engineer Intern
Straggler detection in distributed training
TmaxSoft · System Software Engineer (fulfillment of military obligations)
Education
Sungkyunkwan University
B.Sc., Computer Engineering
Service
Reviewer: NeurIPS 2026, ICML 2026 (Gold Reviewer), IEEE TNET 2023