AI Researcher @ ETRI · Korea University M.S. in AI
Building reliable LLM systems — prompt optimization, hallucination mitigation, on-device deployment
I work on making large language models more reliable under real-world constraints — small models, limited context, fixed decoding, production environments.
- Prompt Optimization — automated, minimal-edit refinement pipelines that lift accuracy without retraining
- Hallucination Mitigation — multi-stage curative approaches to reduce factual errors at inference time
- On-device / sLLM — deploying optimized prompts on resource-constrained, small-scale LLMs
- RAG & Information Retrieval — retrieval-augmented generation for grounded, trustworthy outputs
| 📄 |
POaaS: Minimal-Edit Prompt Optimization as a Service to Lift Accuracy and Cut Hallucinations on On-Device sLLMs EACL 2026 Workshop |
| 📄 |
Multi-stage Prompt Refinement for Mitigating Hallucinations in Large Language Models arXiv 2025 |
| 📄 |
CPR: Mitigating Large Language Model Hallucinations with Curative Prompt Refinement IEEE SMC 2024 |
Patent · 치료적 프롬프트 정제를 통해 언어 모델의 환각 오류를 감소시키는 방법, 장치, 및 프로그램 (KR 10-2024-0102366)
| Period | Role | Organization |
|---|---|---|
| 2025.07 – Present | AI Researcher | Electronics and Telecommunications Research Institute (ETRI) |
| 2021.02 – 2023.01 | Web Developer | Mobile Entropy |
| Period | Degree | Institution |
|---|---|---|
| 2023.03 – 2025.02 | M.S. in Artificial Intelligence | Korea University |
| 2015.03 – 2021.02 | B.S. in Computer Science | Chungnam National University |
Thesis: Multi-stage Prompt Refinement for Mitigating Hallucinations in Large Language Models
AI / ML
Engineering