About Sanghyun Park

Security Researcher · AI-driven Pentester

By day I break vehicles at Autocrypt — offensive tests on real cars and ECUs. By night I hunt 0-days in the Linux kernel. My edge: pairing LLMs with fuzzing, a craft built at KAIST's Software Security Lab.

7 Linux kernel 0-days discovered 14 BusyBox 0-days discovered CVE-2026-53239 Linux XFRM UAF · High · CVSS 7.8 361+ Fuzzers cataloged in Fuzzing Survey

Experience

1. Jun 2025 – Present · Seoul, South Korea

   Security Researcher · Autocrypt

   Vehicle Threat Research Lab. I run offensive tests against in-vehicle networks, ECUs, and wireless attack surfaces — and build the automation tooling behind them.

2. Feb 2023 – Feb 2025 · Daejeon, South Korea

   Graduate Researcher · KAIST Software Security Lab

   Fuzzing and LLM-assisted vulnerability reproduction under Prof. Sang Kil Cha. Built LLM1dFuzz; maintained the Fuzzing Survey.

3. Apr 2022 – Feb 2023 · Daejeon, South Korea

   Research Intern · CNU Mobile Distributed Computing Lab

   Architected a distributed IoT storage system on MinIO: GET +347%, PUT +155% across 4 nodes. Led the edge-cloud migration of a power-plant surveillance system — a custom real-time protocol and a FastAPI integration API.

4. Dec 2021 – Feb 2022 · Indianapolis, USA

   Visiting Scholar · Purdue University

   Built a long-range shot-group measurement system with LoRa, YOLOv5, and edge computing. Published at IEEE SAS 2022.

5. Jul 2021 – Dec 2021 · Daejeon, South Korea

   Research Intern · CNU Software Analysis and Testing Lab

   First research role: automated software testing and analysis.

Selected Work

• Linux Kernel Vulnerability Research2026 – present

  Independent audits of Linux kernel subsystems with LLM-assisted analysis, fuzzing, and manual review. Seven 0-days reported and patched upstream via the kernel mailing lists. Writeups in the CVE archive.

• BBXFuzz2026

  LLM-assisted fuzzing of BusyBox applets: 14 0-days, from harness generation to crash triage. Project writeup.

• LLM1dFuzz2023 – 2025

  Master's thesis. An LLM-guided framework that reproduces 1-day vulnerabilities and seeds directed fuzzing. Published at SECAI 2024.

• Fuzzing Survey2023 – 2025

  Maintained fuzzing-survey.org, a genealogy of 361+ fuzzers and the papers behind them.

Publications

• Systematic Bug Reproduction with Large Language Model
  ESORICS Workshop on Security and Artificial Intelligence (SECAI) · 2024 · long version
• ChatFive: Enhancing User Experience in Likert Scale Personality Test through Interactive Conversation with LLM Agents
  ACM Conference on Conversational User Interfaces (CUI) · 2024
• Feasibility of Measuring Shot Group Using LoRa Technology and YOLO V5
  IEEE Sensors Applications Symposium (SAS) · 2022

Education

• KAIST — M.S., Computer Science (School of Computing) 2023 – 2025 · Advisor: Prof. Sang Kil Cha
• Chungnam National University — B.S., Computer Science and Engineering 2017 – 2023 · GPA 4.21/4.5 · Advisor: Prof. Hoon Choi

Contact

Email is fastest: sanghyun.park.cnu@gmail.com. Also on GitHub and LinkedIn — always happy to connect.