Azure Data Scientist Associate Boot Camp

DreamsPlus provides a thorough Azure Data Scientist Associate Boot Camp in Chennai and online that is intended to give you practical experience and get you ready for the Microsoft data science certification.

Syllabus 

• Design and prepare a machine learning solution (20–25%)
• Explore data, and train models (35–40%)
• Prepare a model for deployment (20–25%)
• Deploy and retrain a model (10–15%)

Design and prepare a machine learning solution (20–25%)

Design a machine learning solution

• Ascertain which compute specs are suitable for a training workload.
• Describe the needs for model deployment.
• Decide the model-building or model-training approach to employ.

Manage an Azure Machine Learning workspace

• Establish a workspace for Azure Machine Learning.
• Use developer tools to manage a workspace and facilitate workplace interaction.
• Set up source control with Git integration.
• Establish and oversee registries

Manage data in an Azure Machine Learning workspace

• Choose Azure Storage options.
• register and keep up datastores
• Construct and oversee data assets.

Manage compute for experiments in Azure Machine Learning

• Establish computational targets for training and experimentation.
• Choose a setting for a use case involving machine learning.
• Set up the serverless Spark compute and Azure Synapse Spark pools, among other connected compute resources.
• Track the use of computing resources.

Explore data, and train models (35–40%)

Explore data by using data assets and data stores

• Obtain and manage data when developing interactive applications.
• Manage interactive data using serverless Spark computing and connected Synapse Spark pools.

Create models by using the Azure Machine Learning designer

• Establish a pipeline for training.
• Consume the designer’s data assets.
• Utilize unique code segments in the designer.
• Review the model and incorporate ethical AI principles.

Use automated machine learning to explore optimal models

• For tabular data, use automated machine learning.
• For computer vision, use automated machine learning.
• For natural language processing, employ automated machine learning.
• Choose and comprehend training options, including algorithms and preprocessing.
• Analyze a machine learning run that is automated and consider appropriate AI practices.

Use notebooks for custom model training

• Develop code by using a compute instance.
• Monitor model training using MLflow.
• Assess a model
• Use the Python SDK v2 to train a model.
• To configure a compute instance, use the terminal. 

Tune hyperparameters with Azure Machine Learning

• Choose a technique for sampling.
• Specify the area of interest.
• Describe the main metric.
• Describe your choices for an early termination.

Prepare a model for deployment (20–25%)

Run model training scripts

• Set the parameters for a script’s task run.
• Set up the computer to execute a job.
• Use information from a data asset for a task.
• Create a job using Azure Machine Learning to run a script.
• Log metrics from a task run using MLflow.
• To troubleshoot job run failures, use logs.
• Set up the environment in which a job will operate.
• Set limits for a task.

Implement training pipelines

• Build a pipeline.
• Transfer data between pipeline phases.
• Execute and plan a pipeline.
• Keep an eye on pipeline runs.
• Make unique parts
• Employ pipelines that are based on components.

Manage models in Azure Machine Learning

• Explain the MLflow model’s output.
• Choose a suitable framework for encapsulating a model.
• Evaluate a model using ethical AI concepts.

Deploy and retrain a model (10–15%)

Deploy a model

• Set up the environment for the online deployment.
• Set up the computing for a deployment in batches.
• Install a model on a web page.
• Install a model on a batch destination.
• Examine a deployed online service.
• To begin a batch scoring job, invoke the batch endpoint. 

Apply machine learning operations (MLOps) practices

• Start an Azure Machine Learning job from GitHub or Azure DevOps, for example.
• Retrain the model automatically in response to modifications or additions of fresh data.
• Explain triggers for event-based retraining.