My personal hint would be - if you have absolute zero knowledge in ML field,- start at one particular sub-field called classification and the simplest (IMHO) AI classification algorithm Naive Bayes . Just search for an info about NB in the wild of internet and you'll find a plenty of articles, tutorials and explanations. Also it's very important what is the exact goal which you are trying to achieve ? If it is an understanding (deep?) of ML algorithms, then my proposal suits here. If your goal is just how faster to make use of some ML algo - then just get some available ML library (which are plenty, for example TensorFlow ) and start using it according to documentation. However as I've said, if you want to learn deeply some algo - there is no other way except to implement it yourself.