Changmin Lee

Changmin Lee

AI Research Engineer  Β·  M.S.-Ph.D. @ UNIST

UAI Lab  Β·  Advisor: Prof. Taesik Gong

I work on |

Email GitHub LinkedIn Google Scholar CV

News

2026.05 πŸ“ Paper "From Volume to Value: Preference-Aligned Memory Construction for On-Device RAG" accepted to ICML 2026. Project Page β†—
2025.08 πŸ† Awarded NRF Master's Fellowship (석사과정생연ꡬμž₯λ €κΈˆμ§€μ›).
2025.03 πŸŽ“ Joined UAI Lab at UNIST as a graduate student (M.S.-Ph.D.).
2024.11 πŸ† Outstanding Paper Award at Korean Institute of Information Technology Fall Conference (undergraduate paper competition).

Research

I build AI systems that are personal, efficient, and deployable on the edge.
Personalized AI & RAG
Memory systems that retain what matters to each user β€” selective, preference-aware, compact.
preference alignment memory construction retrieval
EPIC Β· ICML 2026
On-Device AI
Pushing LLM capabilities to consumer devices β€” tight budgets, real users, no cloud.
edge inference efficiency resource-aware
AI Agents & Memory
Compact, reusable memory for GUI agents that act efficiently on mobile devices.
GUI agents action reuse mobile
Ξ”Agent Β· in preparation

About

Experience

UAI Lab
Graduate Researcher
UNIST Β· Ulsan, Korea
2025.03 β€” Present
  • Advisor: Prof. Taesik Gong
  • Research on Personalized AI Β· On-Device AI Β· AI Agents & Memory

Education

UNIST
M.S.-Ph.D. in Computer Science and Engineering
Ulsan, Korea
2025.03 β€” Present
  • GPA: 4.12 / 4.3
Kyungpook National University
B.S. in Computer Science and Engineering
Daegu, Korea
2019.03 β€” 2025.02
  • GPA: 3.83 / 4.3  Β·  Major GPA: 4.07 / 4.3