Deploying and using predictive solutions

When developing the in-database solution and creating it for continuous development (and also deployment), several aspects should be taken into consideration. First of all, the environment where data scientists will be working. You might give them a powerful, standalone server or even allocate proper seats in the cloud. They will need it, especially when training the model. This is extremely important, as you don't want to have your highly-paid statisticians and mathematicians wait for the models to compute and generate. So, enabling the route to a highly scalable CPU and RAM powerful computations is a must. Second to this, you have to get the data there. Whether it's on cloud or on premises, getting data there (and later also, back) should not be overlooked, as this might also be the point where you will lose precious time. And...