Minbeom Kim

Hi! I am a Research Scientist at SK hynix Research, working on Recursive Self-Improvement (RSI) of AI scientists — ensuring that RSI remains safe and controllable, and preventing harm from misalignment in the increasingly capable agents that emerge from it.

I received my Ph.D. from Seoul National University, advised by Prof. Kyomin Jung. During my Ph.D., I developed Distributional Matching RL — preserving diversity and mitigating reward hacking in language model post-training, thereby enabling scalable test-time scaling of AI agents. I was also a student researcher at Google, and interned at LawZero and NAVER Labs Europe.

Please feel free to reach out to me for a discussion here research@minbeom.kim!