LLM Research Engineer @ ESTsoft corp. (July 2024 ~ Present)
NLP specialist focused on solving real-world problems with language models across diverse domains.
- 🧠 Large Language Models: Parameter-efficient fine-tuning and knowledge distillation
- 🤖 AI Agents: Tool-use and multi-step reasoning capabilities
- 🔍 Retrieval-Augmented Generation: Long-context processing for specialized domains
EXTRA: Integrating external knowledge into multimodal hashtag recommendation system (2023)
HS Won, SM Roh, D Kim, MJ Kim, H Kim, KM Kim
2023 IEEE International Conference on Web Services (ICWS), 719-721
P-Distill: Efficient and Effective Prompt Tuning Using Knowledge Distillation (2025)
HS Won, JY Choi, N Zaman, D Aliyeva, KM Kim
Applied Sciences 15 (5), 2420
University student dropout prediction using pretrained language models (2023)
HS Won, MJ Kim, D Kim, HS Kim, KM Kim
Applied Sciences 13 (12), 7073
Large Language Models are Students at Various Levels: Zero-shot Question Difficulty Estimation (2024)
JW Park, SJ Park, HS Won, KM Kim
Findings of the Association for Computational Linguistics: EMNLP 2024, 8157-8177
PETIS: Intent Classification and Slot Filling for Pet Care Services (2024)
N Zaman, SJ Park, HS Won, MJ Kim, HS An, KM Kim
IEEE Access
Multi-stage prompt tuning for political perspective detection in low-resource settings (2023)
KM Kim, M Lee, HS Won, MJ Kim, Y Kim, SK Lee
Applied Sciences 13 (10), 6252
Discovering Time-Varying Public Interest for COVID-19 Case Prediction in South Korea Using Search Engine Queries: Infodemiology Study (2024)
SH Ahn, K Yim, HS Won, KM Kim, DH Jeong
Journal of Medical Internet Research 26, e63476
Contributed repos:
-
gemini-fullstack-langgraph-quickstart – (PR #35)
Corrected typos and grammatical inconsistencies in docstrings and prompts, improving documentation clarity and code readability. -
openai/harmony – (PR #28)
Fixed a duplicated phrase in the safety warning section of the documentation, improving clarity and professionalism.
- 🏛️ Supreme Prosecutors' Office: Legal domain LLM with RAG system for case processing
- 🎬 Automatic Dubbing System: Duration-based translation with phoneme predictions for lip synchronization across 29+ languages
- Duration-aware Neural Machine Translation for cross-lingual dubbing
- Multi-modal AI agents with reasoning and tool integration
- Parameter-efficient fine-tuning for domain-specific applications
- RAG systems for long-context legal and technical document processing
- 📧 Email: [email protected]
- 🎓 Google Scholar: Profile