Machine Learning Operations

As of April of 2023, according to McKinsey, only 15% of businesses’ Machine Learning projects ever succeed. Another study by Gartner found that only 53% of Artificial Intelligence projects ever make it from prototype to production. Machine Learning Operations (MLOps) borrows from Development and Operations (DevOps) principles to successfully take machine learning models to production. The outcome is higher software quality, faster patching and releases, and higher customer satisfaction. By the end of this course, students will have a strong foundation in Machine Learning Operations (MLOps) and the skills that are necessary to build and deploy Machine Learning models in real-world settings.

Section One: Machine Learning

Section one introduces core concepts and tooling needed to work on any machine learning problem. Students get a practical introduction to the core concepts of Machine Learning, along with the ability to apply them to the most common datasets. They will see key Machine Learning modeling methods, including regression and tree-based methods. They will learn how to measure fairness and bias and follow best practices for responsible Machine Learning and Artificial Intelligence.

Section One Curriculum:

Gain hands-on experience with the Jupyter Notebook application and SageMaker, which are two powerful tools to develop and deploy Machine Learning models. 

Learn about AutoML and AutoGluon, and how to use them to automate the tedious parts of the Machine Learning pipeline, freeing up time for more important tasks. 

Understand the importance of data preprocessing and feature engineering, and how to perform each step effectively. 

Learn about different types of Machine Learning models, including tree-based models, regression models, and ensembling models, and how to select and evaluate the best      model for a given task. 

Understand overfitting and underfitting, and how to avoid them by using regularization techniques. 

Train, tune, test, and evaluate common Machine Learning models, and learn how to use hyperparameter tuning to achieve better performance. 

Section Two: Deep Learning

Section two will explore and demonstrate several key neural network design patterns that are used to create models for Machine Learning problems that rely on text and image data. A combination of lecture and coding exercises are used to show how these models are built and how they can be applied to practical Machine Learning problems.

Section Two Curriculum:

Train models on general datasets by implementing a neural network in PyTorch. 

Preprocess text and image data into a more compatible format for Machine Learning model training. 

Implement a recurrent neural network in PyTorch and use it to train a model on textual data. 

Deploy a convolutional neural network (CNN) in PyTorch and use it to train a model on image data. 

List and define various methods to vectorize text, including count vectorization, term frequency-inverse document frequency (TF-IDF), and word embeddings. 

Use transfer learning to more efficiently and effectively train models on small datasets. 

And more.

Section Three: Internet of Things

Section three will explore Machine Learning at the edge. Students will be introduced to Software-Defined Wide Area Networks, Multi Access Edge Computing, Near Edge / Far Edge technologies and more.