Intro

Self-Evolving Agents & Multimodal AI에 관심 있는 학부생 연구자

KAIST DSAIL에서 AI agents의 self-improvement mechanisms를 연구하고 있습니다. 사람들과 아이디어를 나누고 함께 문제를 풀어나가는 과정을 좋아합니다.

SKKU Systems Management Engineering · KAIST DSAIL Summer Undergraduate Research Intern (2026.06–08)

Email: [email protected] · GitHub: https://github.com/each-yi·

LinkedIn: https://www.linkedin.com/in/inyourzero/

Profile

Education

• 성균관대학교(SKKU) | 시스템경영공학(산업공학) 학사 (2023.02–현재) · GPA 3.86/4.5
• DTU (Technical University of Denmark) | Computer Science | Exchange Student (2026.01–2026.07)

Research Interests

• Agentic AI / Self-Evolving Agents
• Multimodal AI (Vision-Language Models) & multimodal reasoning
• Time-series modeling / anomaly detection

Relevant Coursework

• Introduction to Machine Learning, Introduction to AI
• Database, Applied Statistics 1–2
• Data Structures, Operations Research

Technical Skills

• Python, PyTorch, Pandas, NumPy, scikit-learn
• Model analysis: SHAP, permutation importance
• Tooling: Git/GitHub
• Research Practices: paper reproduction, ablation studies, experimental analysis

Research Experience

KAIST DSAIL — Summer Undergraduate Research Intern (2026.06–08)

• 지도교수: 박찬영 교수님
• Research topic: Self-Evolving AI Agents
• Focus: LLM/VLM agents의 reasoning 및 self-improvement mechanisms 탐구 (e.g., iterative self-training, RL-based reasoning, recursive self-improvement)
• What I did: 경량 베이스라인 재현 및 ablation/분석 실험 구성·수행

URP-I — 의료비 지출 위험과 가계 자산배분 (2025.08–2025.12)

• Ayyagari & He (2017) 재현: 의료비 지출 위험 변화가 자산배분에 미치는 영향
• Methods: Difference-in-Differences(DiD), Fixed Effects, placebo/robustness tests, STATA
• Result: 제도 도입 후 고령층 위험자산 보유 확률 약 +2%p

Selected Projects

1) KAIST–POSTECH–UNIST AI & Data Science Competition — FEM 데이터 기반 이상탐지