We sit down with Erik to discuss gradient descents and how understanding this can lead to demystifying machine learning. We talk about the value of knowing how things work under the covers, as well as maths, functional programming, and many other topics, including why he chose Kotlin, and what he likes and dislikes of the language.