Hyesung Jeon

I am a Ph.D. student in Electrical and Computer Engineering at Seoul National University, advised by Prof. Jae-Joon Kim.

My research focuses on efficient algorithms and serving systems for generative models and AI agents, spanning model compression, PEFT, and efficient attention mechanisms.

I am open to discussions about research collaborations and job opportunities :)

Email | CV | LinkedIn | Github | Scholar

Research Interests

Keywords: Efficient AI Serving, Model Compression, PEFT, KV Cache Compression, Agentic AI Systems

My primary research focus is building efficient algorithms and serving systems for generative models, such as large language models, large multi-modal models, and diffusion models. My work mainly spans the post-training stack, from quantization-aware fine-tuning to KV cache management of agentic LLMs. I am drawn to the emerging efficiency challenges with long-context inference and agentic workloads.

In particular, my research interests lie in:

Model compression (quantization and pruning) and parameter-efficient fine-tuning
KV cache compression and efficient attention mechanisms
Efficient inference systems for LLM-based agents
Hardware-software co-design for low-precision training and inference

Education

Seoul National University Ph.D. Student in Electrical and Computer Engineering Advisor: Prof. Jae-Joon Kim	Mar. 2023 – Present
Seoul National University B.S. in Electrical and Computer Engineering Summa Cum Laude (GPA 4.04/4.30)	Mar. 2019 – Feb. 2023

Publications * Equal contribution

	[7] LRAgent: Efficient KV Cache Sharing for Multi-LoRA LLM Agents Hyesung Jeon, Hyeongju Ha, Jae-Joon Kim ICML, July 2026 Paper \| Code \| Bibtex
	[6] QWHA: Quantization-Aware Walsh-Hadamard Adaptation for Parameter-Efficient Fine-Tuning on Large Language Models Hyesung Jeon, Seojune Lee, Beomseok Kang, Yulhwa Kim, Jae-Joon Kim ICLR, April 2026 Paper \| Code \| Bibtex
	[5] PRESTE: Preserving Tiny Exponent Precision for Efficient Sub-8-bit LLM Inference and Fine-Tuning Jehun Lee, Hyesung Jeon, Juchan Lee, Jae-Joon Kim Preprint, Jan. 2026 Paper \| Bibtex
	[4] L4Q: Parameter Efficient Quantization-Aware Fine-Tuning on Large Language Models Hyesung Jeon, Yulhwa Kim, Jae-Joon Kim ACL, Jul. 2025 Paper \| Bibtex
	[3] Leveraging Early-Stage Robustness in Diffusion Models for Efficient and High-Quality Image Synthesis Yulhwa Kim, Dongwon Jo, Hyesung Jeon, Taesu Kim, Daehyun Ahn, Hyungjun Kim, Jae-Joon Kim NeurIPS, Dec. 2023 Paper \| Bibtex
	[2] Squeezing Large-Scaling Diffusion Models for Mobile Jiwoong Choi, Minkyu Kim, Daehyun Ahn, Taesu Kim, Yulhwa Kim, Dongwon Jo, Hyesung Jeon, Jae-Joon Kim, Hyungjun Kim ICML Workshop on Challenges of Deploying Generative AI, Jul. 2023 Paper \| Bibtex
	[1] AND-Net Based Multi Precision Neural Network Accelerator Design Hyesung Jeon, Jae-Joon Kim B.S. Graduate Thesis (in Korean), Feb. 2023 Paper \| Bibtex

Work Experiences

	MangoBoost (Internship) RDMA System Architecture Design for Data Processing Unit Mentor: Prof. Jangwoo Kim Apr. 2022 – Jun. 2022
	SK Hynix Solution Center (Internship) Deep Learning Network Design on SoC-NAND Validation The Encouragement Prize in the Internship Workshop Mentor: Dr. Yong Lee Jun. 2021 – Aug. 2021 Slides
	Seoul National University, Grad. School of Convergence Science and Technology (Student Researcher) Deep Learning Network Architecture Design for Nano-optical Layer Mentor: Prof. Changsoon Kim Jan. 2021 – Feb. 2021 Code \| Slides
	Seoul National University, Dept. of System Semiconductor Engineering for AI (Student Researcher) Development of English Subtitle Generation Server Development of Visitor Counter Application with Face Detection Model Mentor: Dr. Seokkyu Kwon Sep. 2020 – Feb. 2021 Code (Video Translation) \| Code (Face Counting) \| Slides

Honors & Awards

Qualcomm Innovation Fellowship Korea, Winner L4Q: Parameter Efficient Quantization-Aware Fine-Tuning on Large Language Models	2025
Samsung SAIT Computer Engineering Challenge, 3rd Prize vLLM-based LLM Inference Acceleration on Multi-GPU Systems	2023
SNU Social Responsibility+ Competition, The Encouragement Prize Arduino-based Refrigerator Storage Assistance Agentic System Design	2020
SNU Electrical Circuit Design Workshop, The Excellence Prize	2020
Korea Presidential Science Scholarship, Full Tuition Scholarship	2019 – 2023
Intel International Science and Engineering Fair, Finalist	2018

Skills

Language	English, Korean
Programming Language	Python, C/C++
Deep Learning Frameworks	PyTorch, vLLM, SGLang, CUDA, Triton, Verilog, HLS

Academic Services

Invited Talks

Efficient AI Meetup Korea (2024, 2026)
Common Features of KV Cache Compression in LLMs

Reviewer & Committee

ICLR 2026, NeurIPS 2026

Leadership & Volunteering

SNU Tomorrow's Engineers Membership (STEM), Chairman (2021 – 2023)
Organized and participated undergraduate mentorship and scholarship programs.

Bibtex copied!