Azure Machine Learning Services

Describing Azure Machine Learning Service

• AzureML is a cloud service offered by Azure and hosted by Azure
• AzureML is a cloud service used for building and deploying models
• We can build models using Python code in our local IDE
• We can write and execute Jupyter notebooks in our browser once signed into AzureML
• We can deploy models using Azure DevOps
• AzureML typically is used for model deployment

Describing Azure Machine Learning Studio

• It is a cloud service offered by Azure and hosted by Azure
• It is a cloud service used for building and deploying models
• It can be used for building models within a GUI in our browser
• We can write and execute Jupyter notebooks in our browser
• We can deploy models using Azure DevOps
• AzureML Studio is used for GUI-based model deployment

Describing Azure Databricks

• It is a cloud service offered by Databricks and hosted by Azure
• It is a cloud service used for building and deploying models
• We can build models using Python and R code in our local IDE
• We can write and execute Jupyter notebooks in our browser once signed into Azure Databricks
• We can deploy models using Azure MLFlow

• Azure Databricks is typically used for model analysis

  • Since they provide the compute resources related to data preprocessing or modeling
  • Processsing is handled by Spark

Describing Azure HDInsight

• It is a cloud service offered by Azure and hosted by Azure
• It is a cloud service used for setting up cloud-based Hadoop and Spark clusters for computation
• These Hadoop and Spark clusters are used for processing data via MapReduce or Spark jobs
• Azure HDInsight is not a data store

• However, we can use another Azure service for storing any data processed or collected from our clusters

  • Azure Data Lake
  • Azure Data Factory
  • etc.

• Of all Azure's services, HDInsight is the closest to an IaaS

  • Since, there is some amount of cluster management involved

Describing Azure SQL Server ML Services

• It is a cloud service offered by Azure and hosted by Azure
• It is an add-on for SQL Server
• It is used for building models
• It allows us to execute Python and R code on your local SQL Server environment
• This eliminates the need to switch between the database and machine learning environments

References

• Azure Data Science Certification
• Defining MLOps with Python
• Describing Model Registrey and Deployment
• Describing Pipelines in Azure
• Describing Azure HDInsight
• Comparing Azure HDInsight and Databricks