CV
Basics
| Full name | Jeewoo Sul (설지우) |
| Summary | AI Research Engineer | Specializing in LLM & Agents |
| jeewoo1025@gmail.com |
Work
-
2024.02 - Present Seoul, Korea
Senior Research Engineer
AI Lab, CTO Division, LG Electronics
- Long-term Memory for AI Agents
- Memory/Agents Benchmark & Evaluation
- AI Coding Agent
Education
Publications
-
2024.06.18 Is Prompt Transfer Always Effective? An Empirical Study of Prompt Transfer for Question Answering
NAACL 2024 (Poster) · Minji Jung*, Soyeon Park*, Jeewoo Sul*, and Yong Suk Choi (*equal contribution, alphabetical order)
Investigating the transferability of soft prompts based on features (ex. answer formats, train dataset size) in Question Answering
-
2023.07.15 Balancing Lexical and Semantic Quality in Abstractive Summarization
ACL 2023 (Oral) · Jeewoo Sul and Yong Suk Choi
Proposing a novel training method in which a re-ranker balances the lexical and semantic quality in Abstractive Summarization.
Honors & Awards
- 2025
Grand Prize, LG SW Bootcamp
LG Electronics
1st of 49 teams (194 participants)
- 2022–2023
Brain Korea 21 (BK21) Fellowship
National Research Foundation of Korea
- 2017–2021
Academic Excellence Scholarship
Kwangwoon University
5 semesters
- 2017, 2021
Dean's List Award
College of Software and Convergence, Kwangwoon University
Top 5 by GPA
Invited Talks
- 2023.12
Recent Advances in Abstractive Summarization
Top Conference Session, Korea Software Congress (KSC 2023)
Academic Services
-
Conference Reviewer: EACL 2027 (Industry Track)
-
Workshop Reviewer: REALM Workshop @ EMNLP 2026
Skills
| Model Training/Serving | |
| PyTorch | |
| vLLM | |
| Docker | |
| Slurm | |
| wandb |
| Agent/RAG | |
| LangGraph | |
| LangFuse | |
| FAISS | |
| Turbovec |
| Web/App | |
| FastAPI | |
| Streamlit |
| Programming Languages | |
| Python | |
| Shell | |
| SQL | |
| JS/TS | |
| C |
| Research Areas | |
| LLMs | |
| Agents | |
| Memory and Context Management | |
| Self-Evolution | |
| Agent Skills | |
| Personalization |