Hi! I’m a postdoctoral researcher at the Dynamical Inference Lab, Helmholtz Munich, hosted by Prof. Steffen Schneider. I received my Ph.D. from KAIST AI, advised by Prof. Jaegul Choo.
My goal is to build AI systems that are testable, explainable, and controllable, so that people can understand, trust, and interact with them more naturally. I apply these methods across scientific domains, from medical imaging to language, bridging interpretability research with real-world use.
Previously, I worked with collaborators at NAVER AI Lab, Qualcomm AI Research, and CMU — see my CV for details.
News
- Jun 2026 Poster presentation at the MCML Munich AI Day 2026.
- Jun 2026 ReDesign accepted at ECCV 2026.
- Jun 2026 Talk at HAICON 2026: Human-in-the-loop Concept Discovery and Curation in Vision Foundation Models, on PatchSAE and ConceptScope — video.
- May 2026 Serving as Area Chair for the ICML 2026 Workshop on Mechanistic Interpretability.
- Mar 2026 Started as a postdoctoral researcher at Helmholtz Munich, in the Dynamical Inference Lab with Steffen Schneider.
- Feb 2026 VisualScratchpad accepted to the ICLR 2026 Workshop on Trustworthy AI.
- Nov 2025 Defended my Ph.D. at KAIST AI, advised by Jaegul Choo — thesis: Trustworthy AI in the Wild: Robustness and Interpretability for Enhanced Reliability.
- Sep 2025 ConceptScope accepted at NeurIPS 2025.
- Jun 2025 Invited talk at the Max Planck Institute for Security and Privacy — Trustworthy AI in the Wild.
- Jun 2025 CytoSAE accepted at MICCAI 2025.
- Jan 2025 PatchSAE accepted at ICLR 2025.
- Oct 2024 Started a research internship at NAVER AI Lab (Backbone Research), Seongnam.
- Sep 2024 CaRot accepted at NeurIPS 2024.
- Jul 2024 Joined the Dynamical Inference Lab at Helmholtz Munich as a visiting researcher, supported by an NRF/DAAD fellowship.
- Apr 2024 U.S. patent granted: Adapting machine learning models for domain-shifted data.
- Aug 2023 Started as a visiting scholar at Carnegie Mellon University, School of Computer Science.
- Feb 2022 Started an internship at Qualcomm AI Research, Seoul.
Publications
Selected publications. For the full list, see my Google Scholar.
ICLR 2026 Trustworthy AI WorkshopVisualScratchpad: Grounding Visual Concepts in Large Vision Language Models
Grounding visual concepts in large vision-language models via a attention-based linking mechanism.
NeurIPS 2025ConceptScope: Characterizing Dataset Bias via Disentangled Visual Concepts
Characterizes dataset bias by disentangling visual concepts learned by sparse autoencoders on vision models.
MICCAI 2025CytoSAE: Interpretable Cell Embeddings for Hematology
Sparse autoencoders for interpretable single-cell embeddings in hematology.
ICLR 2025Sparse Autoencoders Reveal Selective Remapping of Visual Concepts During Adaptation
Sparse autoencoders expose how visual concepts are selectively remapped when vision models adapt to new domains.
NeurIPS 2024Towards Calibrated Robust Fine-Tuning of Vision-Language Models
Calibrated, robust fine-tuning method for vision-language models that preserves uncertainty estimates under distribution shift.
ICLR 2023TTN: A Domain-Shift Aware Batch Normalization in Test-Time Adaptation
Domain-shift aware batch normalization layer that improves test-time adaptation under distribution shift.
EMNLP 2021AVocaDo: Strategy for Adapting Vocabulary to Downstream Domain
A vocabulary adaptation strategy for pretrained language models targeting downstream domains.
The easiest way to reach me is by email.