Seoyoon Chae

Hello 👋 My name is Seoyoon Chae, and I am a M.S. student working on efficient and private deep learning systems.

I am currently M.S. student in Electrical and Computer Engineering at Sungkyunkwan University, advised by Prof. Yulhwa Kim.
Prior to that, I received my B.S. from Sungkyunkwan University, Suwon, South Korea.

My research focuses on making large-scale AI models more memory-efficient, deployable, and practical for real-world systems.

🔎 My current interests include:

• Model compression (quantization & sparsity)
• Resource-efficient inference for large language models (LLMs)
• KV-cache optimization for long-context decoding
• Hardware-aware algorithm design (FPGA / ASIC accelerators)
• Efficient dataflow and PIM architectures

📄 Download CV (as of Sep. 2026)

🎯 Research Vision

Modern AI models are becoming increasingly large and computationally demanding.
I am interested in bridging the gap between algorithm design and hardware constraints — designing methods that are not only theoretically sound, but also efficient in memory, latency, and energy.

My long-term goal is to build scalable AI systems that are both intelligent, efficient and reliable.

🧪 Current Research

Resource-Efficient LLM Inference

• KV-cache quantization and adaptive precision scheduling
• Tiling-based memory management strategies
• Eviction of KV-Cache based on geometric features for reasoning tasks

Hardware-Aware Deep Learning

• FPGA-based LLM accelerators: Network Quantization and Verilog Implementation
• RTL Implementation and Verification, Synthesis and Place-and-Route for tape-out of CNN accelerators
• Design of DRAM-based PIM simulator for dynamic-precision LLM inference

📰 News

• July 2026 - I attended the ICML conference 2026.

🎓 Education

Sungkyunkwan University
M.S. in Electrical and Computer Engineering (Sep. 2026 – Present)

Sungkyunkwan University
B.S. in Electronic and Electrical Engineering (Feb. 2022 - Aug. 2026)

📬 Contact

✉️ Email: sychaeee@g.skku.edu