My old colleague Karl Rosaen is taking time off from working to study Machine Learning. He’s meticulously documenting his curriculum and day to day work, good resources there.
One big part of Karl’s studying is Raschka’s book Python Machine Learning. It looks like a nicely practical book, I was looking for something like that last year and it wasn’t quite out yet. His example code is online. And his blog post about writing it is nice.
My biggest regret from my machine learning work is not applying it to real problems. Karl mentions that the Kaggle competitions are a good set of problems and datasets to work on.