This is where we apply the statistical and numerical techniques on the data coming out of data engineering phase to find the best way for the machine to predict the desired outcomes. It includes building models, comparing model performances, choosing the right model / models and deploying them. If required, we also work on the model fairness, accountability and transparency.