Model architecture diagrams are my page-turners. I occasionally fall head over heels for a clever fusion layer, an elegant loss function, or a beautifully designed encoder.
Yu Leng
Yu is a PhD student at the Hertie AI and an IMPRS-IS scholar, working on multimodal AI for patient stratification and outcome prediction in brain health. She is excited about representation learning that integrates heterogeneous data, such as electronic health records (EHRs), longitudinal cohorts, imaging, tissue, and other modalities, into joint patient embeddings. Previously, she was a data scientist at Massachusetts General Hospital, working on clinical NLP/LLM for EHRs and foundation models for digital pathology. Outside academia, Yu is a classical Chinese and ballet dancer. She also enjoys knitting, climbing, and figure skating, and keeps a growing perfume collection.
Model architecture diagrams are my page-turners. I occasionally fall head over heels for a clever fusion layer, an elegant loss function, or a beautifully designed encoder.
Present Positions And Title
PhD Student
Research Group
Email Address
Career
| Period | Institution | Role |
|---|---|---|
| Since 2026 | IMPRS-IS / Hertie Institute for AI in Brain Health, University of Tübingen | PhD Student |
| 2024-2026 | Massachusetts General Hospita | Data Scientist |
| 2023 | Merck & Co. | Data Science Intern |
| 2020-2023 | Roche Pharmaceuticals | Clinical Data Management Intern |
Academic Education
| Year | Degree | Institution | Field of Study |
|---|---|---|---|
| 2024 | M.Sc. | Harvard University | Biomedical Informatics |
| 2022 | B.Sc. | Duke Kunshan / Duke University | Data Science |