Biography
I am an AI researcher working on medical imaging and cardiovascular medicine, currently an Instructor / Faculty Member in the Department of Radiology at Harvard Medical School. My research develops AI methodologies for medical imaging and applies them to real clinical problems.
My primary research areas are AI for echocardiography and cardiovascular image analysis. I have developed a range of AI models spanning the full echocardiographic workflow — from view identification and cardiac structure segmentation to quantitative analysis and disease diagnosis. In particular, I have pursued self-supervised learning and foundation model research on large-scale echocardiography data, resulting in EchoFM, an echocardiography foundation model that leverages the intrinsic periodicity of the heart as a learning signal.
More recently, I have extended my work to coronary imaging and intravascular OCT. Together with the MGH OCT Lab, I use AI to analyze the morphology and microstructure of coronary plaque and to quantitatively assess high-risk lesions.
The ultimate goal of my research goes beyond building high-performing models: I aim to create cardiovascular AI systems that are trustworthy across diverse patients and clinical settings, and that can genuinely be used in everyday clinical practice.
Research Interests
- Medical imaging computing
- Cardiovascular imaging
- Multimodal AI
- Medical foundation models
- Physical AI
- Echocardiography
- Intravascular OCT
At a Glance
Source: Google Scholar.
Education & Appointments
- 2026 –Co-Founder & Chief Technology OfficerEMAI
- 2026 –Instructor / Faculty Member, RadiologyHarvard Medical School
- 2022 – 2026Research FellowCenter for Advanced Medical Computing and Analysis (CAMCA), Department of Radiology, MGH & Harvard Medical School
- 2022PhD, Biomedical EngineeringYonsei University, South Korea · Thesis: Deep Learning on Multi-physical Features and Hemodynamic Modeling for Abdominal Aortic Aneurysm Growth Prediction
- 2019Visiting ResearcherDepartment of Mechanical Engineering, Michigan State University
- 2016Visiting ResearcherDepartment of Radiology, Cedars-Sinai Medical Center, California
- 2016B.E., Biomedical EngineeringYonsei University, South Korea
Contact
- Emailskim207@mgh.harvard.edu
- Office399 Revolution Drive, 11W48.22Somerville, MA 02145
News
- 2026Co-founded EMAI, serving as Chief Technology Officer.
- 2026.01Appointed Instructor (faculty rank) in Radiology at Harvard Medical School.
- 2026Invited talk: Advancing Precision Medicine through Multimodal AI in Cardiology, Yonsei University, South Korea.
- 2026Two abstracts accepted at the ESC Congress 2026 — a mixture-of-experts model for vulnerable plaque characterization, and a contrastive multi-modal foundation model for coronary plaque phenotyping.
- 2026Paper accepted at AAAI 2026: Beyond Adapter Retrieval — Latent Geometry-Preserving Composition via Sparse Task Projection.
- 2026Coronary plaque studies published in JACC: Advances, JACC: Cardiovascular Imaging, and European Heart Journal – Cardiovascular Imaging.
- 2025EchoFM, an echocardiography foundation model, published in IEEE Transactions on Medical Imaging.
- 2025MediViSTA published in IEEE Journal of Biomedical and Health Informatics.
- 2025Three papers accepted at MICCAI 2025 (SAMed-2, BioSAM-2, Cascaded 3D Diffusion) and one at CLIP 2025.
- 2025ECHOPulse accepted at ICLR 2025; fairness-aware segmentation work accepted at ICML 2025.
- 2025Featured in People Who Illuminate Korea, BRIC.
- 2024.11Started the project Imaging and Biomarker-Driven OCT Foundation Model for Comprehensive Plaque Representation and Diagnosis with Ik-Kyung Jang, MD, Harvard Medical School.
- 2024Awarded the Sejong Science Fellowship, Ministry of Science and ICT (MSIT), South Korea.
- 2022.06Joined the Center for Advanced Medical Computing and Analysis (CAMCA), Massachusetts General Hospital and Harvard Medical School, as a Research Fellow.
- 2022.03Completed PhD in Biomedical Engineering at Yonsei University; received the Best Innovative Research Paper award for the doctoral dissertation.