Today Tom and Steven are at the QXday 2018. Tom did the keynote on predictive quality. One of the first things you can do is build a defect predictor yourself! Use the knowledge and tutorials of online environments like Google Tensorflow, learn and train an algorithm with your own defect database. We used the open available defect list of the linux kernel and trained with this data to get a defect predictor that has 72% accuracy. I’m curious to see how you do!