Projects

Research projects in NLP, Information Retrieval, and RAG.

WBL: World Best LLM Project   HuggingFace

  • Led the data team and established a query clarity evaluation framework using engineered GPT-5.2 prompts, training a Qwen3-4B model as a proprietary tagger to selectively curate datasets with diverse clarity levels.
  • Established a robust response filtering pipeline for off-policy SFT data by ensembling three distinct reward models, applying score fusion techniques to accurately evaluate and retain high-quality responses.
  • Curated and refined large-scale alignment samples by integrating the score fusion pipeline with rigorous LLM-as-a-judge and Code Execution metrics to effectively eliminate repetitive and low-quality responses.

KURE: Korea University Retrieval Embedding Model   GitHub | HuggingFace

  • Led the lab’s flagship Korean retrieval project; curated large-scale training datasets and trained a dense retriever that achieved State-of-the-Art (1st place) on the MTEB-ko-retrieval leaderboard (as of Aug. 2025).
  • Designed and maintained MTEB-ko-retrieval, establishing a comprehensive evaluation suite and standardized public leaderboard for the Korean IR community.
  • Open-sourced the framework, achieving 200+ GitHub stars and 2.3M+ cumulative downloads on Hugging Face.
  • Awarded Best Oral Presentation at HCLT 2025.

Korean Sparse & Multi-Vector Retrievers   HF

  • Trained and open-sourced Korean Sparse and Multi-Vector models, achieving State-of-the-Art performance among corresponding architectures (as of Feb. 2026) on the Korean Retrieval Benchmark.
  • Outperformed existing multilingual and Korean fine-tuned models, providing highly optimized, reproducible pipelines to advance dense-sparse hybrid retrieval experiments.

ReviewSearch   Blog | GitHub | Demo

  • Built and deployed a search engine over 205,988 real peer reviews and rebuttals for agents, so that a rebuttal draft is grounded in how authors actually answered the same concern.
  • Trained a hybrid retriever for the domain: a fine-tuned dense encoder and a sparse encoder fused with Reciprocal Rank Fusion.
  • Shipped it as the grounding layer for RebuttalDraft in the rebuttal-skills agent plugin.

URACLE-Korea University Collaborative Research

  • Trained a Korean-English cross-lingual retriever and analyzed language-pair trade-offs; used model merging to recover mono-lingual retrieval while retaining CLIR gains. [Paper]
  • Developed an end-to-end training recipe for a Korean legal-domain LLM; defined real-world judicial tasks, curated expert-written and synthetic alignment datasets, and optimized training strategies to preserve general capabilities.
  • This research directly contributed to KT winning a $10.42 million contract to build an AI platform for the South Korean Supreme Court. [News]
  • Published the training methodologies and data pipelines as LEGALMIDM at the ICLR 2026 Data-FM Workshop.