About Me

I am a Ph.D. candidate at Machine Learning and Intelligence Lab (MLILAB) in Korea Advanced Institute of Science and Technology (KAIST), advised by Prof.Eunho Yang.

My research broadly focuses on large language models, with particular interests in reinforcement learning with verifiable rewards and agentic reinforcement learning.

Publications

Pruning and Distilling Mixture-of-Experts into Dense Language Models
Junhyuck Kim, Jihun Yun, Haechan Kim, Gyeongman Kim, Joonghyun Bae, Jaewoong Cho
Preprint
KVoiceBench, KOpenAudioBench, and KMMAU: Agent-Driven Korean Speech Benchmarks for Evaluating SpeechLMs
Haechan Kim, Seungjun Chung, Inkyu Park, Jihoo Lee, Jonghyun Lee
Preprint
Raon-Speech Technical Report
Beomsoo Kim, Changho Choi, Dohyun Kim, Dongki Lee, Ethan Ewer, Eunchong Kim, Gyeongman Kim, Haechan Kim, Hyeonghwan Kim, Inkyu Park, Jihun Yun, Jihwan Moon, Jiyun Kim, Joonghyun Bae, Junhyuck Kim, Minkyu Kim, Sehun Lee, Seungjun Chung, Sungwoo Cho, Dongmin Park, Dongwon Kim, Hara Kang, Jonghyun Lee, Keon Lee, Kangwook Lee, Jaewoong Cho
Technical Report
Discounted Beta-Bernoulli Reward Estimation for Sample-Efficient Reinforcement Learning with Verifiable Rewards
Haechan Kim, Soohyun Ryu, Gyouk Chu, Doohyuk Jang, Eunho Yang
ICML 2026
LearnerVoice A Dataset of Non-Native English Learners’ Spontaneous Speech
Haechan Kim, Junho Myung, Seoyoung Kim, Sungpah Lee, Dongyeop Kang, Juho Kim
INTERSPEECH 2024, Shortlisted for ISCA Best Student Paper Award 2024

Work Experiences

KRAFTON

Nov. 2025 - Jun. 2026
AI Research Div. · Research Intern · Seoul, South Korea
Advisor: Kangwook Lee

I developed Raon-Speech-9B and Raon-SpeechChat-9B, contributing especially to model probing and evaluation. I also conducted research on agentic reinforcement learning and harnesses, and investigated MoE merging and pruning methods.

KAIST Interaction Lab

Mar. 2023 - Mar. 2024
Visiting Researcher · Daejeon, South Korea
Advisor: Juho Kim

I published a paper titled 「LearnerVoice: A Dataset of Non-Native English Learners’ Spontaneous Speech」, accepted to INTERSPEECH 2024. Additionally, I filed two patents: “System for Learning English Speaking and Method Thereof” and “System for Diagnosing and Learning Pronunciation and Method Thereof”.

Ringle

Jul. 2022 - Mar. 2024
Research Engineer · Seoul, South Korea

I spearheaded the planning, development, and deployment of an AI-driven engine, CAF, which predicts IELTS and TOEFL speaking scores by analyzing conversations between tutors and learners.

Projects

KRAFTON developed A.X K1 to integrate Non-thinking and Thinking modes into a single model, enabling either fast responses or deeper reasoning through a mode-specific chat template. We explored multiple data recipes and found paired training on the same prompts effective for making the model follow the think token rather than surface-level prompt cues. This recipe was incorporated into A.X K1 training.

Blog