After working on multiple projects covering important machine learning concepts, techniques, and widely used algorithms, we have gathered a broad picture of the machine learning ecosystem, as well as solid experience in tackling practical problems using machine learning algorithms and Python. However, there will be issues once we start working on projects from scratch in the real world. This chapter aims to get us ready for it with 21 best practices to follow throughout the entire machine learning solution workflow.

We will cover the following topics in this chapter:

• Machine learning solution workflow
• Tasks in the data preparation stage
• Tasks in the training sets generation stage
• Tasks in the algorithm training, evaluation, and selection stage
• Tasks in the system deployment and monitoring stage
• Best practices in the data preparation stage
• Best practices in the training sets generation stage
• Word embedding
• Best practices in the model training, evaluation, and selection stage
• Best practices in the system deployment and monitoring stage