MLOps Explained

In this blog, we will learn about MLOps.


MLOps was launched for Azure at Microsoft Build 2019. It is known as DevOps for Machine Learning and empowers data scientists and app developers to help bring ML models to production. It is able to audit, certify and re-use assets in the ML lifecycle.
Azure ML is composed of web interfaces and an SDK that trains and deploys a model.
The process of AI creation involves working with lots of data, cleaning the data, writing and running experiments, publishing models, and finally collecting real-world data and improving your models.

Benefits of MLOPS

In a normal application, the developer creates the app in an editor such as Visual Studio. They must commit, code on a develop branch and create a pull request to merge code to the master branch.
The example below shows the app development process 
With Azure ML and AzureDevOps we can manage datasets, experiments, and models.

Model Training

AML provides several ways to train models, from code in the SDK, to the visual designer.

Packaging the Model

In this step, we create an interface configuration.

Validating the Model

Another frequent challenge data scientists face is guaranteeing that models will perform as expected once they are deployed to the cloud or the edge. With new model validation and profiling capabilities, you can provide sample input queries to your model.

Deploying the model

Here, we deploy the model to Azure Container Instances, which offers a simple way to run a container in Azure without the need to provision any virtual machines or adopt a higher-level service.

Monitoring the Model

Audit the model and test performance
Azure DevOps provides the common tools data scientists use to manage code, work items, and CI/CD pipelines. With the Azure DevOps extension for machine learning, we are introducing new capabilities that make it easies to manage ML CI/CD pipelines with the same tools used for the software development processes.
The extension includes the abilities to trigger Azure Pipelines release on model registration, easily connect an Azure Machine Learning Workspace to an Azure DevOps project, and perform a series of tasks designed to help interaction with Azure Machine Learning as easily as possible from the existing automation tooling.