This talk is aimed at developers who want to use machine learning to solve their own binary (2 class) classification task. No prior machine learning or math experience is required. This talk will cover feature engineering (including a robust solution to 'the problem of null data'), predicting the right class with a Random Forest, cross-validating to avoid over-fitting, diagnosing problems in the classifier and approaches to deploying the classifier in the real world. My goal is to provide you with a process that you can take back to the office to try with your own data. It'll be backed by reproducible working code.