Seongryong Jung at EMNLP 2025 Oral

@EMNLP 2025 Oral

Seongryong Jung

I am a research intern at Princeton University under the supervision of Prof. Zhuang Liu. I am also an undergraduate student in Artificial Intelligence at Chung-Ang University in Seoul, Korea. My research aims to build capable AI systems that improve through experience. I study how post-training can help models learn from their interactions and continually strengthen their capabilities, while remaining effective and efficient in real-world use.

At Trillion Labs, I worked on foundation model post-training, with a focus on knowledge distillation and continual learning for large language models. My prior work was nominated for an outstanding paper award at EMNLP.

Email  /  CV  /  Scholar  /  Github  /  LinkedIn

Seeking Ph.D. opportunities starting in Fall 2027!

News

  • Oct. 2026 APSD is now available as a preprint
  • Oct. 2026 Started a research internship at Princeton University
  • May. 2026 Awarded a research grant for the Knowledge Distillation for Self-Evolving AI project (2026 - 2027)
  • Mar. 2026 Started a research internship at Trillion Labs (Mar. 2026 - Jun. 2026)
  • Dec. 2025 Gave an invited talk, Knowledge Distillation for Large Language Models, at KSC 2025
  • Nov. 2025 ToDi was nominated for an Outstanding Paper Award at EMNLP 2025 (top 0.9%)
  • Oct. 2025 Started as an exchange student at Tokyo Metropolitan University (Oct. 2025 - Mar. 2026)
  • Aug. 2025 ToDi was accepted to EMNLP 2025 (Oral)

Publications
APSD framework figure Correctness-Dependent Asymmetric Target Distributions for Self-Distillation Policy Optimization
Suwan Yoon*, Seongryong Jung*, Changhee Lee
Preprint
project / arXiv / bibtex
* Equal Contribution
ToDi method overview ToDi: Token-wise Distillation via Fine-Grained Divergence Control
Seongryong Jung, Suwan Yoon, DongGeon Kim, Hwanhee Lee
Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP 2025)
Oral Presentation / Outstanding Paper Award Nominee (top 0.9%)
project / paper / arXiv / code / bibtex

Experiences

Princeton University Oct. 2026 – Present
Princeton, NJ (Remote)
  • Research Intern (Advisor: Prof. Zhuang Liu).
  • Researching agentic vision-language models.
Korea University Jul. 2026 – Oct. 2026
Seoul, Korea
  • Research Collaborator (Advisor: Prof. Changhee Lee).
  • Developed an asymmetric self-distillation method for LLM post-training (APSD).
Trillion Labs Mar. 2026 – Jun. 2026
Seoul, Korea
  • Research Intern.
  • Worked on continual learning and agentic distillation for foundation model post-training (Gravity).
DMT Labs Mar. 2025 – Jun. 2025
Seoul, Korea
  • Research Intern.
  • Designed a novel post-training distillation method for large language models (ToDi).

Education

Chung-Ang University Mar. 2023 – Present
Seoul, Korea
  • B.S. in Artificial Intelligence
  • GPA: 4.34 / 4.50
  • Research Advisor: Prof. Hwanhee Lee
Tokyo Metropolitan University Oct. 2025 – Mar. 2026
Tokyo, Japan
  • Exchange Student, Faculty of Systems Design

Awards & Honors

  • Research Grant, Knowledge Distillation for Self-Evolving AI (2026)
  • Software College Dean's Award, 2025 CAU Engineering Academic Festival (2025)
  • Outstanding Paper Award Nominee, EMNLP 2025 (2025)
  • IITP President's Award, 2025 ICT Internship Outstanding Achievement Awards (2025)
  • Japan Student Services Organization (JASSO) Scholarship (2025)

Languages

  • Korean, Native
  • English, CEFR B2
  • Japanese, CEFR B2

Trajectory

Self-Evolving AI ··· Mar. 2023 B.S. in AI Chung-Ang University Mar. 2025 Research Intern DMT Labs · ToDi Oct. 2025 Exchange Student Tokyo Metropolitan Univ. Nov. 2025 EMNLP 2025 Outstanding Paper Award Finalist Mar. 2026 Research Intern Trillion Labs · Gravity Oct. 2026 Research Intern Princeton University
Mar. 2023B.S. in AIChung-Ang University
Mar. 2025Research InternDMT Labs / ToDi
Oct. 2025Exchange StudentTokyo Metropolitan University
Nov. 2025EMNLP 2025Outstanding Paper Award Finalist
Mar. 2026Research InternTrillion Labs / Gravity
Oct. 2026Research InternPrinceton University
FutureSelf-Evolving AI