I am a Research Fellow @ AI² Lab, KAIST. My research broadly focuses on evaluating and improving multimodal models: building benchmarks and datasets that expose where models fail before deployment, and developing methods to mitigate those failures. I drive the complete model development lifecycle by architecting data pipelines and formulating robust evaluation frameworks, advancing research in domains ranging from safety-critical autonomous driving to general computer vision. Currently, I am working on vision-language-action (VLA) models with reasoning as well as using agents for scenario generation.
I am also quite interested in these domains: agentic AI, synthetic image data generation, and AI in law.
- Evaluating Video LLMs' Understanding of Corner Cases in Autonomous Driving (to be presented at BMVC 2026) · First author
- RADBench: Evaluating Foundation VLMs for Corner Case Description on A Comprehensive Dataset (IEEE IV 2026, Oral) · First author
- ClusterSplat: Semantic Cluster Selection for 3D Visual Grounding in Gaussian Splatting (to be presented at NeurIPS 2026, Spotlight) · Co-author
- IEEE ITSS Korea Academic Presentation Competition — Grand Prize (August 2026)
- POSCO Asia Fellowship (September 2024 – August 2026)
- Samsung Global Hope Scholarship (March 2021 – February 2024)
- Handong Cornerstone Scholarship (March 2020 – February 2024)

