Kernel methods are a class of versatile tools for machine learning, numerical approximation and scientific computing. I this talk we discuss novel advancements of deep and greedy kernel methods and presents several applications thereof.