Curriculum Vitae

Suhyeong Park

Education

Ph.D. Student

Data Mining & Information Systems Lab (DMIS Lab), advised by Prof. Jaewoo Kang.

B.E. in Data Science

Research Experience

Freelance Researcher

AI Research Intern

  • Developed an LLM agent to track clinical trial research.
  • Developed a retrieval-augmented generation chatbot for pediatric emergency queries.

Undergraduate Researcher

Analyzed ultrasound signals and images and participated in research paper seminars.

Undergraduate Researcher

Applied natural language processing to clinical text and participated in research paper seminars.

Publications

Taeyun Roh, Suhyeong Park, Gijin Kim, Wonjune Jang, and Jaewoo Kang†. October 2026. Learning How to Remember: Memory Modality Selection for GUI Agents. arXiv preprint arXiv:2610.17928.

Suhyeong Park, Junha Jung, and Jaewoo Kang†. September 2026. Reason Through the Latent! Making Latent Visual Reasoning Necessary. arXiv preprint arXiv:2609.06746.

Suhyeong Park, Junha Jung, Jungwoo Park, and Jaewoo Kang†. July 2026. Do All Visual Tokens Matter Equally? Object-Evidence Preserving Token Merging for Vision-Language Retrieval. arXiv preprint arXiv:2607.04605.

Taeyun Roh, Suhyeong Park, Dongyoung Lee, Wonjune Jang, Junha Jung, Eunyeong Jo, and Jaewoo Kang†. March 2026. BUZZY: Contrastive Scoring to Mitigate Text-Induced Bias in Multimodal Multiple-Choice QA. arXiv preprint arXiv:2603.28026.

Su-Hyeong Park*, Ho-Beom Kim*, Seong-Jin Park, Dinara Aliyeva, and Kang-Min Kim†. November 2025. Leveraging Knowledge Graph-Enhanced LLMs for Context-Aware Medical Consultation. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 30459–30475, Suzhou, China. Association for Computational Linguistics.

* Equal contribution † Corresponding author

Projects

Efficient Lesion Search in Ultrasound Images via RL

Trained a PPO agent to learn efficient ultrasound scan paths for lesion localization.

LLM-Based Educational Chatbot with Celebrity Persona

Used web search to model celebrity speech patterns and semantic retrieval to surface curated coding problems during tutoring.

Named Entity Recognition for Kiosk Interaction

Fine-tuned RoBERTa for store classification and extraction of order details such as menu items and quantities.

MPRAG for Open-Domain Question Answering

Proposed momentum updates for the passage encoder and a graph reranker that selects passages using Monte Carlo search.

Additional projects

  • Patient outcome prediction from EHR time series with LSTMs.
  • Whole-slide image classification with attention-based multiple instance learning.
  • Context-aware insurance recommendations with retrieval-augmented generation.
  • Ultrasound speckle reduction using synthesized MRI data for supervision.
  • Financial forecasting with BERT-based sentiment analysis and LSTM time-series modeling.

Additional Experience

NAVER AI RUSH Ambassador

Developed a HyperCLOVA X chatbot that retrieved learning materials, explained them, and generated summary notes. Shared example uses through outreach.

Executive Member

Led an NLP paper review seminar and represented the club in the U.S. Embassy in Korea’s Tech Talk Program.

Honors & Awards

2024 Biohealth Challenge

First prize (team) for applying machine learning to Doppler ultrasound signals for breast cancer detection.

2023 Data Analysis Competition

First prize (team) for detecting anomalies in printed circuit board images.

2023 Data Analysis Competition (Medical AI Track)

First prize (team) for classifying whole-slide pathology images with multiple instance learning.

2022 Data Analysis Competition

First prize (team) for classifying pet diseases from structured veterinary data.

Technical Skills

Programming

Python, C, C++

Machine Learning and Data Analysis

Hugging Face, LangChain, LangGraph, PyTorch, TensorFlow, scikit-learn, NumPy, pandas, Matplotlib, Seaborn

Edit CV

CV entry details