This is a zero to hero course. We will start with no (or little) knowledge about machine learning and end the day with creating our custom APT detection that matches commercial grade tools.

Along the way, we will look at different types of machine learning, explore their limitations, and discuss typical problems. We will answer relevant questions, like how to identify that the machine has really learned something useful, how to deal with resource constraints, perform feature selection, and how to fine tune towards our goal.

The whole course is filled with real-world applications: After playing with artificial data, we will create a classifier for a polymorphic malware family, and end the day with your own threat prevention AI that uses features from ClamAV.