Younghan Park

I am Younghan Park, a master's student at Carnegie Mellon University's Machine Learning Department. I got my BS in CSE from Yonsei University, where I had a fortune to be advised by Prof. Youngjae Yu. Also, I spent my junior year at The University of Texas at Austin, where I was fortunate to be advised by Prof. Eunsol Choi and Prof. David Harwath.

My long-term goal is to build machines (including robots!) that can perceive the world and understand/use language in a human-like way. My current areas of interest are:

Multimodal Language Understanding: How can we teach machines to understand the information that arises from the fusion of modalities? What if the context is complex, long, or unstructured?
Embodied Language Grounding: How can we ground natural language in a machine's perception and action, so that it understands and carries out human instructions as intended?

Email: younghanpark [at] yonsei [dot] ac [dot] kr
Curriculum Vitae / GitHub / Google Scholar / LinkedIn

News

(May 2026) One first-authored paper was accepted to Interspeech 2026, and one co-authored paper was accepted to ICML 2026 Workshop.
(Feb 2026) I will be joining Carnegie Mellon University for my master's study starting Fall 2026.
(Dec 2025) I started my internship at KRAFTON Inc., as a deep learning engineer intern in the AI Research Division.
(Oct 2025) I completed my internship at NAVER Cloud Corporation.
(Aug 2025) I graduated with high honors at Yonsei University.
(Jul 2025) I have been selected as a recipient of the Korean Government Scholarship Program for Study Overseas (대한민국 정부 국비유학).
(Jul 2025) My first first-authored paper is accepted at COLM 2025.
(Jul 2024) I am excited to share that I have been selected as a recipient of the Presidential Science Scholarship (대통령과학장학금).

Education

Carnegie Mellon University , Pittsburgh, PA, United States MS in Machine Learning, School of Computer Science, Aug 2026 - Dec 2027 (Expected)
Yonsei University , Seoul, South Korea BS in Computer Science and Engineering, College of Computing, Mar 2020 - Aug 2025
The University of Texas at Austin , Austin, TX, United States Exchange Student (Major: Computer Science), Aug 2023 - May 2024

Honors and Awards

The Korean Government Scholarship Program for Study Overseas, Starting Fall 2026
$80,000 in pursuit of master's degree in US (selected as one of three recipients in the digital category)
Presidential Science Scholarship, Mar 2024 - Aug 2025
Full tuition and stipend for junior and senior years, awarded in the name of the President of the Republic of Korea (selected as one of 60 students nationwide)
Graduation with High Honors, Yonsei University, Aug 2025
Awarded to 2 out of 88 students in the College of Computing
High Honors, Yonsei University, Fall 2024
Merit-based Scholarship, Yonsei University, Spring 2021, Spring 2023, Fall 2023
University Honors, The University of Texas at Austin, Fall 2023
Honors, Yonsei University, Fall 2020, Spring 2021, Spring 2023
Best Tutor, Yonsei University, Spring 2021
Yonsei CSI2100: Computer Programming

Publications and Patents

*: equal contribution, ^: equal advising

Preprints / Under Review
	Phonemizing User-Generated Text: A Benchmark, Taxonomy, and Compositional Approach Minju Jeon, Younghan Park, Han Sung Park, Jong-Hwan Kim, Dong-Jin Kim, Hoyeon Lee Under Review at Empirical Methods in Natural Language Processing (EMNLP), 2026 paper (coming soon)
Conference Publications / Workshop Papers
	LLM-Based Multi-Reference Evaluation for Efficient and Robust Assessment of Phrase Break Annotations Younghan Park, Hoyeon Lee, Hawon Jeong, Jong-Hwan Kim Interspeech, 2026 paper (coming soon)
	AgentVidBench: A Multi-Hop Video Question Answering Benchmark for Evaluating MLLM Agents Seoyeon An, Hyeonseo Jang, Minsu Kim, Chanho Lee, Younghan Park, Kangwook Lee Workshop on Failure Modes of Agentic AI @ International Conference on Machine Learning (FAGEN@ICML)*, 2026 paper (coming soon)
	Rhapsody: A Dataset for Highlight Detection in Podcasts Younghan Park, Anuj Diwan, David Harwath^, Eunsol Choi^ Conference on Language Modeling (COLM), 2025 paper / poster / code / dataset
Patents
	Control Apparatus using Oral Cavity (구강을 이용한 조종장치) Kyuhyun Hwang, Gideok Jang, Woonchae Jeong, Jaein Yoo, Younghan Park, Yonsoo Kim Korea Patent Registration No. 10-2020-0140424. Granted. Google Patents / The Seoul Economic Daily / The Digital Times / The Korea Economic Daily

Experience

KRAFTON Inc. (The Publisher of PUBG: Battlegrounds), Dec 2025 - Present Intern, AI Applied Research Dept., AI Research Division
Naver Cloud Corporation , Jun 2025 - Oct 2025 Research Intern, Voice Technology Team
NLP Group , The University of Texas at Austin, Aug 2023 - May 2025 Visiting Student (Advisors: Prof. Eunsol Choi, Prof. David Harwath)
Multimodal Intelligence Research Lab , Yonsei University, Feb 2023 - Aug 2023 Undergraduate Research Intern (Advisor: Prof. Youngjae Yu)
Korean Augmentation to the US Army (KATUSA) , Jul 2021 - Jan 2023 Specialty: Military Police, Final Rank: Sergeant
MouthMouse , Apr 2020 - Jul 2021 Co-Founder, Software Engineer

Selected Projects

	RL Agent for Multi-turn Negotiation in Chat Conversations , Spring 2024 report / demo (YouTube) / code Built a multi-turn negotiation agent by fine-tuning an LLM to propose multiple candidate utterances and buy prices per turn. Trained offline RL agents from model response representations.
	Robust Speaker Diarization under Real-world Noise , Fall 2023 report Built a speaker diarization pipeline and evaluated 11 preprocessing configurations, including source separation and speech enhancement, under synthetic background noise and crosstalk.
	CO2Gather , Jan 2022 - Feb 2022 project page / code An NLP software analyzes card logs to automatically calculate potential savings, fostering sustainable collaboration between consumers, businesses, and governments. Top 2 Winner, KPMG Ideathon Challenge 2022 National Round, KPMG Korea
	MaGam , Jan 2021 - Feb 2021 project page / code A computer vision-based full-stack web application for highly accurate mask detection by addressing the challenge of limited data using both classical and modern features. Top 3 Winner, KPMG Ideathon Challenge 2021 National Round, KPMG Korea