Using Azure DevOps to collaborate Data Scientists and App Developers

What is your most favorite IDE? A developer can mention various of IDE’s. How about data scientist?

Normally, the Developers develop the code and data scientists train develop and train the model. Tools such as VSCode come in handy as easy to install and use. Many have many preferences so the chances are that it can be a choice of the developer or the development team. Usually the trained model is handed over to the app developer for integrating it and build the final application. There are times where the mismatches in compatibility can cost both the app developer and the model developer. The resulting friction between app developers and data scientists to identify and fix the root cause can be a slow, frustrating, and expensive process

We often here organizations including managers continuously talking about Artificial Intelligence. People like to find solutions that are integrated with AI. So as the developers have a development lifecycle, the data scientists follow a data science lifecycle. 

The lifecycle includes processes such as,

Data Ingestion --> Data Preparation --> Model Development --> Model Deployment

There can be many iterations of this lifecycle as there can be requirements for changing the data labels, removing anomalies, changes upon user feedback and timely decision changes and many more. 

Serverless with Azure Kubernetes Service

As I have said before, Kubernetes is the future and organizations are migrating to containers. That makes Azure a top choice in the industry. Even non container based solutions right now are planned with intentions to get into cloud in future. 

Speed, reliability and portability are among the main reasons for people to move to containers. Those qualities make virtual machines a less preferred option. Self healing reduces maintenance overlook needs.

DevOps Integration with GitHub and Azure

Customers need faster, yet reliable solutions. Developers would love to relax the weekends after doing a deployment. Both these did not happen together very often. As a result, various iterative development methodologies such as Agile and Scrum are introduced. DevOps is the latest explanation introduced by Microsoft for such organizations where the team of developers and the requirement holders are connected very efficiently.

Microsoft made GitHub a friend by acquiring GitHub rather than sticking only with Visual Studio Teams services. Thanks to that decision, over 10,000 open source projects are handled with the support of both GitHub and Azure DevOps TOGETHER.

Visual interface for Azure Machine Learning Service

Microsoft has been great with interfaces for visualizing workflows. SharePoint has it, Microsoft Flow has it and many other services including Logic Apps has it. The preview for the Azure ML visual interface was announced during Microsoft Build. 

Not only visualizing and understanding, drag and drop helps ease the processes of testing and deploying ML models as well. Nothing helps more than a visual diagram for a developer to understand the logic of, may be his own work sometime later. 

Source: Azure Blog