Welcome to Vantage 101, the machine learning engineering course that bridges the gap between university and the business world. This course is designed to quickly get a grasp on all best practices you wish you would have known at the beginning of your career.
Vantage 101 is the course a Machine Learning Engineer wishes they had done at the beginning of their career. Many early-career data scientists and related specialists fall into a skills gap where their theoretical knowledge far exceeds their ability to apply it. Usually the theoretical knowledge is obtained in highly-structured environments such as a university degree, whereas the knowledge on how to apply tends to be much less structured and accessible. Often, someone who falls into that gap will feel a slowdown in their productivity, their satisfaction with their work, and their career progression.
We believe that gap is avoidable, and we try to bridge it with Vantage 101 - a structured introduction to Machine Learning Engineering (MLE). The course seeks to provide a hands-on introduction to every critical aspect of MLE, focusing on gaining "must have" knowledge and applying it immediately, while delegating the "nice to have" knowledge to curated external resources. The course also includes a case study, which gives you a platform to apply what you learned in each chapter's assignment.
In short, this course will guide you as you go from a Jupyter Notebook with a trained machine learning model to a production model deployed according to industry best practices.
We want to treat this course as a living thing that keeps evolving with the industry, its audience and its creators. We welcome feedback and suggestions, especially recommendations for high-quality external resources on topics we cover, as more eyes are simply better at finding more gold nuggets online.
Vantage 101 is intended for those with a Data Science/Statistics/Mathematics or similar background who want to learn or refresh a baseline knowledge in Machine Learning Engineering. Therefore, the ability to train a machine learning model in a Jupyter Notebook is assumed and we would recommend coming back to this course later to those that do not have this yet.
The aim is to concisely teach the skills, best practices, and tools for going from a Jupyter Notebook to a deployed production model. Therefore, you can expect of every chapter that it:
The code and other files for the case can be found in this Github repository. Chapter 1 will explain the case and the file structure, and is the best starting point for following the rest of the course. Following the sequence of chapters is critical until after chapter 2: Setting up your local machine, but after that the order is simply our recommendation.
We believe that dry theory is not enough to fully grasp a concept, which is why we also use an example case to apply all our methods on. Throughout the course we will be using the Australian rainfall case. We will use the dataset and a basic notebook with a model to start off our course. Again, everything you need to get started with the case can be found in this Github repository, and you will see plenty of code snippets throughout the chapters as well.