About me
I am a Ph.D. candidate at MIT EECS & LIDS, advised by Asu Ozdaglar. I am closely working with Kaiqing Zhang. My research experience includes previous work with Amazon AGI and NAVER AI (industry), as well as with the Ryu Optimization Group and the Bioinformatics Group (academia).
My current research interest is multi-agent learning:
- Single/Multi-agent Sequential Decision Making LLMs Training with provable guarantees.– It should combine game theory, sequential decision-making theory, and understandings of the recent advances in VLM.
Feel free to reach me! I actively discuss various topics with folks! I am open to collaboration (but not advising currently), so please don’t hesitate to reach out if you’re excited about these topics and would like to work together!
I love games!!!! Life itself is a game :)
I have held a certification as a Korean actuary since 2018.
I run ML-Theory-Korea with my colleagues. Feel free to email me (cpark97 at mit.edu) to join ML-theory-Korea. We are actively researching several areas.
Mentee
- Hwangyu Jeong (Seoul National University): March 2025 - . Working on MDM Theory
- Valentio Iversion (University of Waterloo): June 2024 - . Working on LLM Theory
Former Mentee
- Jaeyeon Kim (SNU, Currently Harvard CS): Oct 2022 - Oct 2023. Worked on Optimization Theory (H-duality). Now, he is my great co-worker.
News
- Sep 2025: I am honored to be selected as a Amazon AI Research Innovation Fellow.
- June 2025: BehaviorSFT: Behavioral Token Conditioning for Clinical Agents Across the Proactivity Spectrum paper got accepted to EMNLP 2025 as a finding.
- May 2025: Multi-agent RL for LLMs paper got accepted to ACL 2025 as a main paper, and also has been accepted for an oral presentation for ICML 2025 Workshop: Multi-Agent Systems in the Era of Foundation Models: Opportunities, Challenges and Futures.
- Feb 2025: Check out the Multi-agent RL for LLMs paper. I’m really enthusiastic about its framework, which enables the training of multi-agent LLM systems.
- Jan 2025: LLM-no-regret and Blackswan-Hypothesis are accepted for ICLR 2025. Blackswan-Hypothesis is also accepted for an oral paper for ICLR 2025 Financial AI Workshop. LLM-no-regret has been accepted for an oral presentation for ICLR 2024 AGI Workshop: How Far Are We From AGI.
- October 2024: A paper on Adaptive-Multi-Agent-for-MedAI has been accepted for an oral presentation for NeurIPS 2024.
- September 2024: I am honored to be selected as a Takeda Fellowship Recipient, which supports students conducting research at the intersection of artificial intelligence and health.
- February 2024: A paper on Personalized-RLHF is acceppted at CVPR 2024. I contributed by providing efficient RLHF with groupwise comparison and its theoretical analysis. Also, a paper on Inequality-to-optimal-method is accepted at JMLR. This work was done when I was an undergraduate intern at SNU.
- February 2024 I got a master’s degree!
- November 2023: I am honored to be selected as a top reviewer in NeurIPS 2023.
- September 2023: I am honored to be selected as Siebel Scholars Recipient.
- September 2023: Papers on Markov Network Games and H-duality are accepted at NeurIPS 2023.
- September 2022: Paper on a Loss function perspective of mixed sample data augmentation accepted at NeurIPS 2022. This work was done when I was an intern at NAVER AI Lab. I got NeurIPS travel awards!
- September 2022: I started my PhD at MIT! I also received an Xianhong Wu Fellowship from MIT.
Contact
Email: cpark97[at]mit.edu