Youngjoon Kim

I am a postdoctoral researcher at SSLab, School of Cybersecurity and Privacy, Georgia Tech, working with Professor Taesoo Kim. My research focuses on AI-driven automated vulnerability discovery and repair via Cyber Reasoning Systems (CRS), combining LLMs with program analysis techniques.

As a member of Team Atlanta, I won 1st place in the DARPA AI Cyber Challenge (AIxCC) Finals, where our CRS found 43 vulnerabilities including 6 zero-days. I am currently a main contributor to OSS-CRS.

I received my Ph.D. in Information Security from Korea University under the supervision of Professor Jiwon Yoon (Thesis: AI-Driven Automated Vulnerability Discovery and Repair). Previously, I served as a captain in the R.O.K. Army, working at the Agency for Defense Development (ADD) and the R.O.K. Cyber Operations Command, gaining experience in both security research and security engineering.

My main research interest is LLM agent-based offensive security. More broadly, my interests include fuzzing, automated program repair, LLM-based security analysis, and offensive security.

News

Jun 2026	Our paper CtxFuzz was accepted to ASE 2026!
May 2026	Our paper SoK: DARPA’s AI Cyber Challenge was accepted to USENIX Security 2026!
Apr 2026	Our paper OSS-CRS was accepted to USENIX WOOT 2026!
Aug 2025	Team Atlanta won 1st place in the DARPA AI Cyber Challenge (AIxCC) Finals! Our CRS found 43 vulnerabilities including 6 zero-days.
May 2025	Our paper “Logs In, Patches Out: Automated Vulnerability Repair via Tree-of-Thought LLM Analysis” was accepted to USENIX Security 2025!

Publications

ASE

CtxFuzz: Context-Aware Directed Fuzzing via Runtime Fuzzer–Agent Cooperation

Jiho Kim, Dongkwan Kim, HyungSeok Han, and 6 more authors

In IEEE/ACM International Conference on Automated Software Engineering (ASE 26). To appear , 2026

Bib

@inproceedings{kim2026ctxfuzz,
  title = {CtxFuzz: Context-Aware Directed Fuzzing via Runtime Fuzzer–Agent Cooperation},
  author = {Kim, Jiho and Kim, Dongkwan and Han, HyungSeok and Kim, Youngjoon and Park, Soyeon and Ji, Sangwoo and Wang, Joshua and Jeong, Dae R. and Kim, Taesoo},
  booktitle = {IEEE/ACM International Conference on Automated Software Engineering (ASE 26)},
  year = {2026},
}

SEC

SoK: DARPA’s AI Cyber Challenge (AIxCC): Competition Design, Architectures, and Lessons Learned

Cen Zhang, Younggi Park, Fabian Fleischer, and 18 more authors

In 35th USENIX Security Symposium (USENIX Security 26). To appear , 2026

Bib PDF

@inproceedings{zhang2026sok,
  title = {SoK: DARPA's AI Cyber Challenge (AIxCC): Competition Design, Architectures, and Lessons Learned},
  author = {Zhang, Cen and Park, Younggi and Fleischer, Fabian and Fu, Yu-Fu and Kim, Jiho and Kim, Dongkwan and Kim, Youngjoon and Xu, Qingxiao and Chin, Andrew and Sheng, Ze and Zhao, Hanqing and Pelican, Michael and Musliner, David J. and Huang, Jeff and Silliman, Jon and Mcdaniel, Mikel and Casavant, Jefferson and Goldthwaite, Isaac and Vidovich, Nicholas and Lehman, Matthew and Kim, Taesoo},
  booktitle = {35th USENIX Security Symposium (USENIX Security 26)},
  year = {2026},
}

WOOT

OSS-CRS: Liberating AIxCC Cyber Reasoning Systems for Real-World Open-Source Security

Andrew Chin, Dongkwan Kim, Yu-Fu Fu, and 7 more authors

In 20th USENIX WOOT Conference on Offensive Technologies (WOOT 26). To appear , 2026

Bib PDF Code

@inproceedings{chin2026osscrs,
  title = {OSS-CRS: Liberating AIxCC Cyber Reasoning Systems for Real-World Open-Source Security},
  author = {Chin, Andrew and Kim, Dongkwan and Fu, Yu-Fu and Fleischer, Fabian and Kim, Youngjoon and Han, HyungSeok and Zhang, Cen and Lee, Brian Junekyu and Zhao, Hanqing and Kim, Taesoo},
  booktitle = {20th USENIX WOOT Conference on Offensive Technologies (WOOT 26)},
  year = {2026},
}

SEC

Logs In, Patches Out: Automated Vulnerability Repair via Tree-of-Thought LLM Analysis

Youngjoon Kim, Sunguk Shin, Hyoungshick Kim, and 1 more author

In 34th USENIX Security Symposium (USENIX Security 25). Hyoungshick Kim and Jiwon Yoon are corresponding authors , 2025

Bib PDF

@inproceedings{kim2025logs,
  title = {Logs In, Patches Out: Automated Vulnerability Repair via Tree-of-Thought LLM Analysis},
  author = {Kim, Youngjoon and Shin, Sunguk and Kim, Hyoungshick and Yoon, Jiwon},
  booktitle = {34th USENIX Security Symposium (USENIX Security 25)},
  year = {2025},
}

ICLR

Enhancing Graph of Thought: Enhancing Prompts with LLM Rationales and Dynamic Temperature Control

Sunguk Shin and Youngjoon Kim

In International Conference on Learning Representations (ICLR). Youngjoon Kim is the corresponding author , 2025

Bib PDF

@inproceedings{shin2025enhancing,
  title = {Enhancing Graph of Thought: Enhancing Prompts with LLM Rationales and Dynamic Temperature Control},
  author = {Shin, Sunguk and Kim, Youngjoon},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2025},
}

More publications →