Hongseung Yu
M.S. student at Seoul National University
I am Hongseung Yu, a M.S. candidate in Electrical and Computer Engineering (ECE) at Seoul National University, advised by Prof. Kyunghan Lee.
My research centers on efficient computing systems for machine learning, with a focus on LLM inference. I am particularly interested in designing inference systems that are tightly coupled with the underlying environment—hardware and operating systems. My interests also extend to hardware-software co-design and resource-constrained edge computing.
Publications
-
mzCache: On-Device LLM Memory Management under Multitasking
Hongseung Yu, Minsung Kim, Jongseok Park, Kyunghan Lee
Accepted to ACM MobiCom 2026, Austin, TX, USA
On-device LLM inference systems are subject to OS memory management because mobile SoCs use a unified memory architecture. As users switch between apps, model weights and KV cache are evicted under multitasking pressure, and restoring them on the next request significantly delays the response. mzCache introduces a memory management scheme designed around the underlying hardware resources and the characteristics of LLM memory, and is the first LLM inference system to manage LLM memory itself rather than leaving it to the OS. Implemented on llama.cpp and deployed as an Android application, it reduces Time-to-First-Token by 2.1–5.5× compared to storage-backed partial offload.