You'll learn the architecture of an end-to-end ML solution. What is docker and how to use containers; how to split code into modules; how to run modules to create an automated process; how to write classes and functions using clean code techniques.
When it comes to building models, we are always dealing with repetition of complex processes of choosing / learning / tuning / validating / testing an algorithm. And the difficulty lies in the fact that in most cases the result of all this work is a long Python code in a Jupyter notebook. Duh..
Even if model development is the foundation of everything, it is equally important to keep your code clean and structured in simple and interchangeable building blocks. In this lecture I will introduce you to MLOps fundamentals. We will learn how to structure your development, automate modules and avoid impediments using unit tests, config files and versioning.
By the end of this lecture you will understand how a typical architectureand structure of an ML project looks like, and will be able to independently create simple but properly structured process chains in a docker container.
Note that this is not about building models, but rather about the correct architecture to help improve the process overall. If you are not familiar with MLOps I would strongly recommend do some readings first.
To get the most out of this workshop, participants should have some understanding of the Python programming language (classes, functions, sklearn library) and at least a general idea of the machine learning process.
See you there!