CV / Portfolio
Seong-Gyu Park

Seong-Gyu Park

박성규

AI Security Researcher

AI Security Researcher specializing in Trustworthy LLMs and AGI Safety. Dedicated to uncovering and mitigating vulnerabilities in the reasoning processes of Large Language Models (LLMs).

Research AI Security & Trustworthy LLMs
Affiliation Soongsil University
Seoul, Korea parksk99@soongsil.ac.kr
GitHub LinkedIn Google Scholar
— visitors

Experience

Researcher

Mar 2026 - Present

Soongsil University

Committed to building robust, safe, and reliable AI systems for high-stakes environments

Graduate Student

Mar 2024 - Feb 2026

Soongsil University

Committed to building robust, safe, and reliable AI systems for high-stakes environments

Skills

Programming Languages

C / C++ Python

Frameworks & Tools

PyTorch Git / GitHub Linux / Unix

Publications

1

STAR: Detecting Inference-time Backdoors in LLM Reasoning via State-Transition Amplification Ratio.

Seong-Gyu Park, Sohee Park, Jisu Lee, Hyunsik Na, Daeseon Choi.

arXiv:2601.08511 (Submitted to ACL 2026), Jan 2026

View Paper
2

ASTRA: Adversarial Stealthy Trigger Reasoning Attacks for Black-Box LLMs

Seong-Gyu Park, Sohee Park, Daeseon Choi

The 30th Pacific-Asia Conference on Knowledge Discovery and Data Mining(PAKDD 2026), Jun 2026 (To be presented)

View Paper
3

Adversarial Image Detection for Vision Transformers via Attention Map

Seong-Gyu Park, Hyunsik Na, Daeseon Choi

IEEE International Conference on Advanced Visual and Signal-Based Systems (AVSS 2025), Aug 2025

View Paper
4

Motion based user authentification for metaverse.

Seong-Gyu Park, Gwonsang Ryu

Journal of the Korea Institute of Information Security & Cryptology (2024): 493-503., Jan 2025

View Paper

Awards & Honors

🏆

Best Paper Award
(Minister's Award, Ministry of the Interior and Safety (MOIS), Republic of Korea)

Korea Institute of Information Security & Cryptology

2025
🏆

President's Award, IT Software Competition

Soongsil University

2023