The Microsoft Certified: Azure Data Scientist Associate certification is designed for candidates who are proficient in implementing data science as well as machine learning to complete tasks based on Azure. Such candidates are able to create an appropriate working environment for data science workloads based on Azure, carry out data experiments, optimize and deploy machine learning models in production. In addition, exam takers are knowledgeable of data science and possess skills to utilize Azure Machine Learning and Azure Databricks. The certification process includes passing one exam – DP-100.
The Microsoft DP-100 exam is about designing and applying a data science solution based on Azure. Since the exam is intended for professionals, it will last for 100 minutes or 120 minutes. If there are labs included, it will be longer (2 hours). Within the time allocated for this Microsoft evaluation, you have to complete from 40 to 60 questions of various types. Some of them to mention include drag and drop, multiple choice, active screen, best answer, build list, and hot spot. The exam is priced at $165.
Knowing more about the exam, its content is divided into 4 parts. The first one refers to dealing with Azure resources for machine learning. This implies your ability to cover an Azure Machine Learning workspace, managing data there, operating compute to carry out experiments. In addition, your skills to apply access control in Azure Machine Learning, create an environment for Azure Machine Learning development, along with an Azure Databricks workspace will be evaluated. That’s why you should have skills to work with an Azure Machine Learning studio, workspace settings, datasets, and Azure Databricks.
The second topic covers carrying out experiments and training models. This topic measures the examinees’ competency in using the Azure Machine Learning designer, running model training scripts, creating metrics from the performed experiment. You should be able to implement Azure Machine Learning designer to create a training pipeline, utilize custom code modules in designer, apply the Azure Machine Learning SDK through running an experiment, use MLflow to track experiments. In addition, you should demonstrate your skills to design optimal models through using Automated Machine Learning, as well as tune hyperparameters with its help.
The third part is dedicated to deploying and operationalizing machine learning solutions. This topic spins around the following skills: choosing compute for model deployment, using a model as a service, operating models in Azure Machine Learning, and creating an Azure Machine Learning pipeline. Your proficiency in utilizing the Azure Machine Learning SDK in order to apply the pipelines will also be checked. Furthermore, you should be familiar with applying ML Ops practices. This means that you are able to refactor notebooks into scripts, apply their source control, and initiate an Azure Machine Learning pipeline from Azure DevOps.
The major emphasis in the fourth topic is placed on applying responsible machine learning. This topic is built around the skills of implementing model explainers to interpret models, explaining fairness and as well as privacy considerations for models. In this part, you should be ready to demonstrate your skills in specifying a model interpreter, generating feature importance data, assessing model fairness, and explaining principles of differential privacy.
According to the figures stated on the Ziprecruiter website, being a certified professional in Azure Data Scientist you can earn about $122,096 per annum.