Skip to content
View Jaeuk-Han's full-sized avatar
  • Korea, Gyeonggi-do
  • 17:25 (UTC +09:00)

Highlights

  • Pro

Block or report Jaeuk-Han

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Jaeuk-Han/README.md

About

AI & Software @ Gachon University · Korean NLP · Information Retrieval · Human-AI Interaction

Email · LinkedIn


About Me

I am an undergraduate in the Department of AI & Software at Gachon University, interested in building AI systems that are reliable, useful, and grounded in evidence.

My work spans Korean NLP, retrieval-augmented generation, LLM fine-tuning, and human-centered decision support. I have developed dense retrieval and reranking pipelines, applied LoRA and GRPO-based fine-tuning, and built AI applications for language, cultural heritage, legislative analysis, and emergency medicine.

I am especially interested in connecting rigorous experimentation with real-world applications in culture, education, and public services.

Recent Highlights

  • 🏆 Outstanding Trainee, NAVER Cloud CEO Award (우수교육생 · 네이버클라우드 대표이사상), NAVER Cloud Academy Generative AI Advanced Program, 2026
  • ☁️ NAVER Cloud Platform Certified Professional AI (NCP-AI), 2026
  • 🏁 Finalist, 2025 AI Malpyeong Competition, National Institute of Korean Language
  • 🧬 Solo Top 7%, The 2nd Medical AI Competition, Korea University College of Medicine

Selected Work

Built a Korean grammar QA system combining a dual-encoder retriever, hard-negative training, cross-encoder reranking, and GRPO-based generation. Qualified for the final round of the 2025 AI Malpyeong Competition.

Structured 3,500+ recipes from 132 historical texts (886-1947) and developed an evidence-grounded QA pipeline using BM25, dense retrieval, MMR, and cross-encoder reranking.

Voice-based Intelligent Triage & Ambulance Link — Developed a real-time emergency room recommendation system using public medical APIs, hospital resource data, FastAPI, map-based routing, and interpretable pre-KTAS logic.

Built an end-to-end representation-learning pipeline for genomic mutations and conducted systematic ablations across backbones, pooling strategies, and projection heads. Ranked in the top 7% as a solo participant in the 2nd Medical AI Competition.

Experience

  • Research Intern, CONNECT-AI Research Center — Yonsei University Health System · Jul 2026-Present
    Working on LLM-based triage research as part of a government-funded project related to the National Fire Agency.

  • NAVER Cloud Academy — Generative AI Advanced Program · May-Jun 2026
    Completed advanced training in generative AI and received the NAVER Cloud CEO Award as an outstanding trainee.

  • Undergraduate Research Assistant, ISNLP Lab — Gachon University · Aug-Dec 2025
    Developed retrieval-augmented pipelines and conducted experiments in dense retrieval, LoRA, GRPO, and reliability evaluation.

  • AI & Robotics Instructor, Taekwang High School · May-Oct 2024
    Delivered workshops and special lectures on machine learning, deep learning, and Arduino using slides and hands-on notebooks.

What I Use

Languages

Frameworks & Libraries

Tools

Certifications & Recognition

  • NAVER Cloud Platform Certified Professional AI (NCP-AI) · 2026
  • TOPCIT Level 3 · Nov 2025
  • NAVER Cloud CEO Award — Outstanding Trainee · 2026

Open to research collaboration and projects in NLP, retrieval, reliable AI, and human-centered AI systems.

Pinned Loading

  1. korean-traditional-liquor-datasetkorean-traditional-liquor-datasetPublic

    Korean Traditional Liquor Dataset (한국 전통주, Sool/술) — curated from historical brewing texts, including ingredients, steps, and yields. Data is parsed, normalized, and enriched with derived features …

    Python

  2. Korean_QA_RAG_2025Korean_QA_RAG_2025Public

    Korean Grammar QA RAG system with Dual Encoder and GRPO reinforcement learning

    Python

  3. law-demand-dashboardlaw-demand-dashboardPublic

    Topic-labeled news/SNS are matched to Korean legislative bills using Qwen3-Embedding-8B (LoRA-fine-tuned); major bills are selected per topic and presented as a static HTML leaderboard.

    Python

  4. taekwang-ai-lectures-2024taekwang-ai-lectures-2024Public

    Educational materials from the AI & Robotics Workshop Series conducted at Taekwang High School’s “Fourth Industrial Revolution and Mathematics” club in 2024. The 8-session course (May–October, 3 ho…

    Jupyter Notebook

  5. 2025-ermct-project2025-ermct-projectPublic

    Decision-support system for routing emergency patients to appropriate hospitals in Seoul, developed as part of the 2025 Gachon University P-Practical Project. It is built for paramedics to improve …

    Python 2

  6. glm-mutation-embeddingsglm-mutation-embeddingsPublic

    Mutation-sensitive DNA embeddings using dna2vec with staged metric learning for CD/CDD/PCC.

    Python