Research & Projects

Team Capabilities & Research Background

Research capabilities built by the CEO and team through university and national R&D projects. We provide services based on this background.

The projects and papers below are research conducted by the CEO while affiliated with Korea University HIAI (Human-Inspired AI Research) before the company was established.

Major Projects

KURE

2024–

Top 100 National R&DNRFHIAI

Research on truthfulness and consistency of Korean generative AI. Selected as 2025 Top 100 National R&D achievement.

KULLM

2023

Open SourceLLMHIAI

Korean instruction-tuning open-source LLM. One of the first publicly available Korean LLMs in Korea.

View →

MMTEB

2025

ICLR 2025EmbeddingBenchmark

Massive Multilingual Text Embedding Benchmark. Established multilingual text embedding evaluation framework (co-author).

KoE5

2024

EmbeddingDataset

Dataset construction for Korean embedding model training.

Ko-Gemma

2024

LLMFine-tuning

Korean-specialized model fine-tuning based on Google Gemma.

GLiNER-ko

2024

NEROpen Source

Korean named entity recognition (NER) model.

Power Plant RAG

2024

RAGEMNLP 2024 Industry

Retrieval-Augmented Generation research for power plant predictive maintenance systems.

HAITEDU chatbot

2026

AI AgentEducation

Educational AI chatbot service. Deployed at haitedu.com.

View →

Papers

2025
MMTEB: Massive Multilingual Text Embedding Benchmark
ICLR 2025
2024
Intelligent Predictive Maintenance System using RAG
EMNLP 2024 Industry
2024
KoE5: A New Dataset for Korean Embedding
arXiv
2023
KULLM: Learning to Construct Korean Instruction-following Large Language Models
arXiv
2022
Dense-to-Question and Sparse-to-Answer: Hybrid Retrieval for Long-form Open-Domain Question Answering
AACL 2022

Open Source Contributions

KULLMvllmhuggingface chat-uiMTEBGLiNERchainlitkss

Inquire About Collaboration