This chapter gives a high-level overview of MLOps principles and framework components important for deployment and life cycling.

The modern MLOps framework

ML workflows

MLOps benefits

Life-cycling stages

App vs. model

Decommissioning

The model life cycle: recap

MLOps components

Automated sequence

Stores and registries

DevOps or MLOps?

MLOps in a Nutshell

This chapter is dedicated to all the considerations we need to make already in the development phase, in order to ensure a smooth ride when we reach the operations.

Our ultimate goal is to explain how to train the model using MLOps best practices and build a model package that enables smooth deployment, reproducibility and post-deployment monitoring.

Deployment-driven development

Testing your machine learning model

Best time to start deploying

Profiling, versioning, and feature stores

Feature store properties

Profiling and feature store benefits

Ensuring reproducibility

Model build pipelines in CI/CD

Different pipelines

Model build pipeline properties

CI/CD integration

Model packaging

Model formats

Full package

Develop for Deployment

This chapter deals with critical model operations questions such as:
- What are the different ways in which we can serve our models?
- What is an API, and what are its key functionalities?
- How do we thoroughly test our service before making it available to the end users?
- How do we update models in production without service disturbance?

You will learn about batch prediction, real-time prediction, input and output data validation, unit testing, integration testing, canary deployment, and much more.

Serving modes

Offline or online?

When time matters - a bit

Building the API

Client-server

API functionalities

Deployment progression and testing

Which test is it?

Progression through environments

Tests per environment

Model deployment strategies

A fitting deployment strategy

Order of risk

Shadow of the shadow

Deploy and Run

This final chapter is dedicated to monitoring and maintaining ML services after they are deployed, as well as to model governance.

You will cover crucial concepts such as verification latency, covariate shift, concept drift, human-in-the-loop systems, and more.


Monitoring ML services

Options

True

False

Shift vs drift

Latency

Already?

Monitoring and alerting

The monitoring system

Alerting

Model maintenance

Data-centric approach

Model-centric approach

Data-centric vs Model-centric

Human-in-the-Loop

Model governance

In the scope of model governance

Not in the scope of model governance

Elements of governance

Design

Development

Operations

Stages of governance

Risk classification

Wrap up

Monitor and Maintain

Explore the modern MLOps framework, including the lifecycle and deployment of machine learning models. In this course, you’ll learn to write ML code that minimizes technical debt, discover the tools you’ll need to deploy and monitor your models, and examine the different types of environments and analytics you’ll encounter.

After you’ve collected, prepared, and labeled your data, run numerous experiments on different models, and proven your concept with a champion model, it’s time for the next steps. Build. Deploy. Monitor. Maintain. That is the life cycle of your model once it's destined for production. That is the Ops part of MLOps.

This course will show you how to navigate the second chapter of your model's journey to value delivery, setting the benchmark for many more to come.

<h2>MLOps Deployment and LifeCycling</h2>

Explore the modern MLOps framework, including the lifecycle and deployment of machine learning models. In this course, you’ll learn to write ML code that minimizes technical debt, discover the tools you’ll need to deploy and monitor your models, and examine the different types of environments and analytics you’ll encounter.

<h2>Learn About the MLOps Lifecycle</h2>
After you’ve collected, prepared, and labeled your data, run numerous experiments on different models, and proven your concept with a champion model, it’s time for the next steps. Build. Deploy. Monitor. Maintain. That is the life cycle of your model once it's destined for production. That is the Ops part of MLOps.

This course will show you how to navigate the second chapter of your model's journey to value delivery, setting the benchmark for many more to come. You’ll start by exploring the MLOps lifecycle, discovering the importance of MLOps and the key functional components for model development, deployment, monitoring, and maintenance.
<h2>Develop ML Code for Deployment</h2>
Next, you’ll learn how to develop models for deployment and how to write effective ML code, leverage tools, and train ML pipelines. As you progress, you’ll cover how to deploy your models, exploring different deployment environments and when to use them. You’ll also develop strategies for replacing existing production models and examine APIs.
<h2>Learn How to Monitor Your Models</h2>
As you complete the course, you’ll discover the crucial performance metrics behind monitoring and maintaining your ML models. You’ll learn about drift monitoring in production, as well as model feedback, updates, and governance.

By the time you’re finished, you’ll understand how you can use MLOps lifecycle to deploy your own models in production.


MLOps Concepts

In this course, you’ll explore the modern MLOps framework, exploring the lifecycle and deployment of machine learning models. 

MLOps Deployment and LifeCycling

MLOps Deployment and Life Cycling

Engenheiro de machine learning

MLOps Fundamentals

Is it ethical to use Machine Learning in this use case?

Is it legally justified to use personal data in this use case?

How will you detect and eliminate unethical model bias?

Are the experiments during model development tracked and stored?

Is the training data versioned and produced in a transparent manner?

Do we perform continuous model monitoring?

Stages of governance

MLOps Deployment and Life Cycling

Exercício interativo prático