EczemaNet: Automating Detection and Severity Assessment of Atopic Dermatitis

I build probabilistic models of noisy real-world systems to improve decision making. Right now that means forecasting the outcome of sports events; previously it was electricity demand and prices, and before that the progression of eczema severity in patients.

In particular, I developed expertise in Bayesian modelling, time-series forecasting, and decision analysis. I am also interested in the engineering around models, in order to deploy reproducible, maintainable, and well-tested solutions in production.