Michelle Espranita Liman

I am a doctoral researcher at the Chair for AI in Medicine at the Technical University of Munich (TUM), specializing in deep learning for echocardiography. I aim to develop a foundation model for echocardiograms, which would enable various tasks—such as heart chamber segmentation and cardiac risk assessment—without finetuning.
I am also exploring multimodal integration with ECG & text, and tokenization techniques to improve computational efficiency for video-based echocardiograms.
Prior to my doctoral studies, I earned my master’s degree at TUM in 2024. My thesis, conducted in collaboration with Harvard Medical School and Massachusetts General Hospital, focused on recovering cropped thoracic CT scans using diffusion models.

Interests:

• Echocardiography
• Video Understanding
• Video Tokenization
• Multimodal AI

Education: 

• Data Engineering and Analytics (M.Sc.), 2024
  Technical University of Munich, Germany
• Computer Engineering (B.Sc.), 2021
  University of Duisburg-Essen, Germany