First, you’ll learn about the core features of MLOps. You’ll explore the machine learning lifecycle, its phases, and the roles associated with MLOps processes.

What is MLOps?

Explaining MLOps

MLOps principles

Different phases in MLOps

The ML lifecycle

Tasks per phase

Roles in MLOps

Your MLOps team

Core roles in MLOps processes

Introduction to MLOps

Next, you’ll learn about the design and development phase in the machine learning lifecycle. You’ll explore added value estimation, data quality, feature stores, and experiment tracking.

MLOps design

Identifying business requirements

Options

Data Scientist

Subject Matter Expert

Business stakeholder

Key metrics

Data quality and ingestion

Determining data quality

Accuracy

Completeness

Consistency

Data quality dimensions

Feature engineering

Improving model performance

Feature engineering approaches

Helping a friend

Experiment tracking

What can we track?

Skipping experiment tracking

Design and Development

In this chapter, you’ll dive into the concepts relevant to deploying machine learning into production, such as runtime environments, containerization, CI/CD pipelines, and deployment strategies.

Preparing model for deployment

Runtime environments

To containerize or not to containerize

Machine learning deployment architecture

Microservice

API data flow

CI/CD and deployment strategy

CI/CD pipelines

Deployment strategies

Automation and scaling

How components automate and scale

MLOps components

Deploying Machine Learning into Production

Finally, you’ll learn about maintaining machine learning in production, with concepts such as statistical and computational monitoring, retraining, different levels of MLOps maturity, and tools that can be used within the machine learning lifecycle to simplify processes.

Monitoring machine learning models

Machine learning monitoring

Statistical vs. computational monitoring

Retraining a machine learning model

Examples of drift

Retraining

Levels of MLOps maturity

MLOps maturity

MLOps maturity levels

MLOps tools

Using a tool

Pick the right tool

Recap: MLOps concepts

Maintaining Machine Learning in Production

Understanding Machine Learning Operations (MLOps) is essential for any data scientist, engineer, or leader to take machine learning models from a local notebook to a functioning model in production. This course introduces you to the key processes, phases, and levels of MLOps, including design, development, deployment, and monitoring. You'll discover how automation enables organizations to efficiently launch, monitor, and update their machine learning models.

<h2>Learn about Machine Learning Operations (MLOps)</h2>
Understanding MLOps concepts is essential for any data scientist, engineer, or leader to take machine learning models from a local notebook to a functioning model in production.
<br><br>
In this course, you’ll learn what MLOps is, understand the different phases in MLOps processes, and identify different levels of MLOps maturity. After learning about the essential MLOps concepts, you’ll be well-equipped in your journey to implement machine learning continuously, reliably, and efficiently.
<br><br>
<h2>Discover How Machine Learning Can be Scaled and Automated</h2>
How can we scale our machine learning projects using the minimum time and resources? And how can we automate our processes to reduce the need for manual intervention and improve model performance? These are fundamental Machine Learning questions that MLOps provides the answers to.
<br><br>
In this MLOps course, you’ll start by exploring the basics of MLOps, looking at the core features and associated roles. Next, you’ll explore the various phases of the machine learning lifecycle in more detail.
<br><br>
As you progress, you'll also learn about systems and tools to better scale and automate machine learning operations, including feature stores, experiment tracking, CI/CD pipelines, microservices, and containerization. You’ll explore key MLOps concepts, giving you a firmer understanding of their applications.

Understanding Machine Learning

Understanding Data Engineering

Discover how MLOps can take machine learning models from local notebooks to functioning models in production that generate real business value.

MLOps Concepts

Associate AI Engineer for Data Scientists

Machine Learning Engineer

Machine Learning in Production in Python

MLOps Fundamentals

It seems that a lot of customers filled in 123456 for their postal code.

After investigation, it was found that 30% of the values are incorrect.

The age of the customer is missing in 45% of all customers.

For one product group, 60% of values are missing.

In one table, the customer's 'age' is the time since their first order; in another table, it's their actual physical age.

Comparing two tables, we find that the same customer id can refer to two different customers.

Data quality dimensions

MLOps Concepts

Hands-on interactive exercise