Precision-medicine AI Lab · PAI Lab
Building network-based, knowledge-guided, and foundation-model-powered AI systems for computational genetics, pharmacogenomics, electronic health records, and AI-driven drug discovery.
I am an Assistant Professor in the Department of Computational Biology at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), and a Visiting Scholar at Harvard Medical School. I lead the Precision-medicine AI Lab (PAI Lab), where we develop next-generation AI systems for computational genetics, pharmacogenomics, and AI-driven drug discovery.
Our mission is to advance AI for precision medicine by integrating multi-scale biomedical data — including genomics, transcriptomics, electronic health records, biomedical knowledge graphs, foundation models, and large-scale biobank data — to improve disease diagnosis, treatment optimization, and therapeutic discovery.
Previously, I was a Postdoctoral Research Fellow working with Professor Tianxi Cai at Harvard Medical School, and served as a data scientist at the Veterans Affairs Boston Healthcare System. I received my Ph.D. in Computer Science from Zhejiang University in 2020.
🚀 I am looking for postdoctoral fellows, Ph.D. students, master’s students, and visiting researchers working at the intersection of AI and precision medicine. Interested candidates are warmly welcome to send me their CV and research statement.
♠️ Outside research, I enjoy playing poker and ultimate frisbee (飞盘).
Our research focuses on developing AI-driven frameworks, particularly network-based, knowledge-guided, and foundation-model-powered AI models, to advance precision medicine by integrating multimodal biomedical data, including biomedical knowledge graphs, electronic health records (EHRs), genomics, transcriptomics, protein and molecule foundation models, and large-scale biobank data. We aim to understand the complex interplays among medications, genetic variants, molecular profiles, and diseases or phenotypes.
A unifying view of my research: integrating biological networks, transcriptomics, biobank data, and clinical evidence to study variant-to-phenotype interpretation, drug development and repurposing, pharmacogenomics, and EHR-based translational research.
Create robust, interpretable, and generalizable AI methodologies that bridge basic biomedical research and clinical practice.
Understand the clinical phenotypic consequences of genetic variants and their therapeutic or adverse effects on medications.
Leverage bulk and single-cell transcriptomic signatures with biomedical knowledge graphs and EHR-derived evidence to discover drug–gene–disease/ADR relationships, predict context-specific drug synergy, and support drug development, repurposing, and safety assessment.
Develop robust models for incident phenotyping, disease risk prediction, rare disease diagnosis, cancer recurrence estimation, and data-driven medical knowledge graphs.
Full publications: Google Scholar
Journal impact factors use the latest available JCR / publisher metrics where available. * indicates equal contribution; † indicates corresponding or senior contribution where applicable.
Lightweight page-view counter with an approximate continent-level distribution. IP addresses are not displayed or stored by this homepage code.
Counts are page views, not unique visitors. The continent estimate comes from the analytics endpoint and is aggregated only.