Bongjun Jang

I'm a Master's student at KAIST School of Computing. I'm interested in applying formal methods (PL, Logics) and AI to real-world systems. So far, I have focused on compiler research: automatically finding optimization bugs and generating optimization rules with formal guarantees. I sometimes write blog posts on what I have read, my thoughts and experiences.

Contact

Personal Email: malloc099@gmail.com
Work Email: bongjun.jang@kaist.ac.kr
GitHub: bongjunj

Education

M.S., School of Computing, KAIST, Korea (Feb. 2025 – Feb. 2027, expected)
B.S., School of Computing, KAIST, Korea (Feb. 2018 – Feb. 2025)
- Minor in Cybersecurity, New York University (Fall 2023)

Publications

Optimization-Directed Compiler Fuzzing for Continuous Translation Validation (PLDI 2025)
- Jaeseong Kwon, Bongjun Jang, Juneyoung Lee, and Kihong Heo
- [Paper] [Slides] [Talk]

Experiences

Programming Systems Laboratory, KAIST (Jan. 2024 – Present) [Website]
- Advisor: Prof. Kihong Heo
- Focused on compiler research: Translation Validation, Peephole Optimization

Open Source Contributions

Cranelift Compiler: Bytecode Alliance Recognized Contributor (Jan. 2026 – Present) [Roster]
- Contributed more than 800 optimization rules for its mid-end optimizer
- Found a critical performance bug that caused a 66% slowdown in compiled programs [GitHub]
LLVM: Reported 55 Miscompilation Bugs [GitHub]

Teaching

Introduction to Information Security (CS.30408), KAIST (Spring 2025) [Course] – Best TA Award
Program Reasoning (CS.40204), KAIST (Fall 2026) [Course]

Military Service

Korean Augmentation to the United States Army (KATUSA), Republic of Korea Army (Feb. 2021 – Aug. 2022)