Machine learning is the next golden child for defenders, promising to solve all their challenges. Outside of attacking these solutions directly, research applying these new toys to red team challenges is difficult to find. How can red teams collect, analyze, and use the data available to them? What are the practicalities of using ML for red purposes? Can ML actually assist an operator? How about become one? This talk will tackle these questions from the ground up.We'll share code that explores the following concepts: How to start processing and analyzing data, Sandbox detection with decision trees, neural networks, and word embeddings, Inferring AD control relationships with fuzzy logic, Teaching a reinforcement learning algorithm to operate like a human, It's not magic, it's math.