Because both of them are required, so you can't think in one without the other. 1
...
This entails planning and creating a suitable working environment for data science workloads on Azure, running data experiments and training predictive models, managing and optimizing models, and deploying machine learning models into production." 1
1
JamesJalaPayneo