Machine learning gives computers the ability to learn without being explicitly programmed. Machine learning techniques are currently being used in applications such as speech and image recognition, language translation, medical diagnosis and research, recommender systems, user behavior analytics, game playing, and many more. The recent development of powerful and inexpensive cloud computing resources has greatly expanded the use of machine learning techniques to new applications, including wind turbine reliability. This presentation will provide an overview of how machine learning works, and will show real world examples of how machine learning is currently being applied to wind turbine SCADA data to predict failures, improve power production, and extend turbine life.