Andy Man Yeung Tai

I’m Andy Tai, a Postdoctoral Researcher at the Institute for Artificial Intelligence in Medicine (IKIM), University Hospital Essen, and the Kamp Lab at the Lamarr Institute and TU Dortmund University, supervised by Prof. Michael Kamp. I work on federated and privacy-preserving machine learning for medicine: methods that let a clinical model learn from data held at many hospitals without any of those hospitals releasing patient records. The interesting problems are the ones that constraint creates, namely how much the sites have to communicate, what the privacy guarantee actually protects against, and whether a model trained this way holds up when the sites disagree.

My main project is FLIP-IT (Federated Learning in Practice Networks), funded by NEXT.IN.NRW with the European Union and the state of North Rhine-Westphalia. I am building federated risk prediction models for chronic kidney disease across general practice clinics, using the Flower framework, supervised by Prof. Michael Kamp (Lamarr Institute and TU Dortmund University), with Prof. Jens Kleesiek (IKIM) and Dr. Moon Kim (IKIM).

I completed my PhD in Neuroscience at UBC in 2024, applying machine learning to addiction psychiatry. As Co-Investigator on the Health Canada-funded Risk Assessment and Management Platform (RAMP, $1.4M, 2019 to 2024), I built overdose risk prediction models on the BC Provincial Overdose Cohort of 36,679 cases. From 2024 to 2026 I was a Postdoctoral Teaching and Learning Fellow in UBC’s Master of Data Science program, where I taught more than 550 students across ten graduate and undergraduate courses and co-authored The Regression Cookbook: Machine Learning and Stats Flavours.

Outside the university I have consulted on the HAVA maritime criminal detection system with Clause Technology and NCIS, I work as an analyst at NAI Innovations, and I developed concussion subtype analysis at Concussion RX.

Write to me if you want to talk about federated learning, medical AI, or clinical decision support.