Intro

I am a 3rd year PhD student at the Algorithmic Intelligence Lab at KAIST AI. My research focuses on developing efficient and safe methods for enhancing, aligning, and personalizing generative models, with an emphasis on language models. I have broad research interests, including generative models, reinforcement learning, AI safety, and nature-inspired intelligence, among others. I always strive to more deeply understand the fundamental principles behind everything I work on.

Previously, I worked as a senior software engineer at Google, developing machine learning methods to localize Google Assistant for low-resource languages. I also worked in Display Ads, building backend systems and improving auction algorithms for better user experience. I received an MS in Computer Science from Stanford University and a BS in Computer Science with a minor in Applied Mathematics from Cornell University.


Selected Publications

  • Correct Answers from Sound Reasoning: Verifiable Process Supervision for Language Models K. Kim, K. Wang, Y. Xie, P. Xu, P. Sheng, C. Wei, Z. Wang, J. Shin, P. Viswanath, S. Oh. arXiv:2605.12519. Paper
  • Self-Refining Language Model Anonymizers via Adversarial Distillation K. Kim, H. Jeon, J. Shin. NeurIPS 2025. Paper Code
  • Personalized Language Models via Privacy-Preserving Evolutionary Model Merging K. Kim, J. Shin, J. Kim. EMNLP 2025 (oral). Paper Code
  • Mamba Drafters for Speculative Decoding D. Choi, S. Oh, S. Dingliwal, J. Tack, K. Kim, W. Song, S. Kim, I. Han, J. Shin, A. Galstyan, S. Katiyar, S. B. Bodapati. EMNLP 2025 Findings. Paper
  • Learning to Contextualize Web Pages for Enhanced Decision Making by LLM Agents D. Lee, J. Lee, K. Kim, J. Tack, J. Shin, Y. W. Teh, K. Lee. ICLR 2025. Paper Code
  • Optimized Feature Generation for Tabular Data via LLMs with Decision Tree Reasoning J. Nam, K. Kim, S. Oh, J. Tack, J. Kim, J. Shin. NeurIPS 2024. Paper Code
  • Margin Matching Preference Optimization: Enhanced Model Alignment with Granular Feedback K. Kim, A. Seo, H. Liu, J. Shin, K. Lee. EMNLP 2024 Findings. Paper Code
  • Confidence-aware Reward Optimization for Fine-tuning Text-to-Image Models K. Kim, J. Jeong, M. An, M. Ghavamzadeh, K. Dvijotham, J. Shin, K. Lee. ICLR 2024. Paper Data
  • BEHAVIOR-1K: A Benchmark for Embodied AI with 1,000 Everyday Activities and Realistic Simulation C. Li, C. Gokmen, G. Levine, R. Martín-Martín, S. Srivastava, C. Wang, J. Wong, R. Zhang, M. Lingelbach, J. Sun, M. Anvari, M. Hwang, M. Sharma, A. Aydin, D. Bansal, S. Hunter, K. Kim, A. Lou, C. Matthews, I. Villa-Renteria, J. Tang, C. Tang, F. Xia, S. Savarese, H. Gweon, K. Liu, J. Wu, F.-F. Li. CoRL 2022 (oral). Paper Code Data
  • Taskmaster-1: Toward a Realistic and Diverse Dialog Dataset B. Byrne, K. Krishnamoorthi, C. Sankar, A. Neelakantan, D. Duckworth, S. Yavuz, B. Goodrich, A. Dubey, A. Cedilnik, K. Kim. EMNLP-IJCNLP 2019. Paper Data