Exams Vouchers: Designing and Implementing a Data Science Solution on Azure (DP-100)
- Price: Please call
- Code: DP-100
Please call
Jump to:
Description
TopThe Azure Data Scientist applies their knowledge of data science and machine learning to implement and run machine learning workloads on Azure; in particular, using Azure Machine Learning Service. This entails planning and creating a suitable working environment for data science workloads on Azure, running data experiments and training predictive models, managing and optimizing models, and deploying machine learning models into production.
Candidates for this exam apply scientific rigor and data exploration techniques to gain actionable insights and communicate results to stakeholders. Candidates use machine learning techniques to train, evaluate, and deploy models to build AI solutions that satisfy business objectives. Candidates use applications that involve natural language processing, speech, computer vision, and predictive analytics.
Candidates serve as part of a multi-disciplinary team that incorporates ethical, privacy, and governance considerations into the solution. Candidates typically have background in mathematics, statistics, and computer science.
Objectives
Top
Content
TopThe bullets that appear below each of the skills measured are intended to illustrate how we are assessing that skill. This list is not definitive or exhaustive.
Set up an Azure Machine Learning workspace (30-35%)
Create an Azure Machine Learning workspace
- create an Azure Machine Learning workspace
- configure workspace settings
- manage a workspace by using Azure Machine Learning Studio
Manage data objects in an Azure Machine Learning workspace
- register and maintain data stores
- create and manage datasets
Manage experiment compute contexts
- create a compute instance
- determine appropriate compute specifications for a training workload
- create compute targets for experiments and training
Run experiments and train models (25-30%)
Create models by using Azure Machine Learning Designer
- create a training pipeline by using Designer
- ingest data in a Designer pipeline
- use Designer modules to define a pipeline data flow
- use custom code modules in Designer
Run training scripts in an Azure Machine Learning workspace
- create and run an experiment by using the Azure Machine Learning SDK
- consume data from a data store in an experiment by using the Azure Machine Learning SDK
- consume data from a dataset in an experiment by using the Azure Machine Learning SDK
- choose an estimator
Generate metrics from an experiment run
- log metrics from an experiment run
- retrieve and view experiment outputs
- use logs to troubleshoot experiment run errors
Automate the model training process
- create a pipeline by using the SDK
- pass data between steps in a pipeline
- run a pipeline
- monitor pipeline runs
Optimize and manage models (20-25%)
Use Automated ML to create optimal models
- use the Automated ML interface in Studio
- use Automated ML from the Azure ML SDK
- select scaling functions and pre-processing options
- determine algorithms to be searched
- define a primary metric
- get data for an Automated ML run
- retrieve the best model
Use Hyperdrive to rune hyperparameters
- select a sampling method
- define the search space
- define the primary metric
- define early termination options
- find the model that has optimal hyperparameter values
Use model explainers to interpret models
- select a model interpreter
- generate feature importance data
Manage models
- register a trained model
- monitor model history
- monitor data drift
Deploy and consume models (20-25%)
Create production compute targets
- consider security for deployed services
- evaluate compute options for deployment
Deploy a model as a service
- configure deployment settings
- consume a deployed service
- troubleshoot deployment container issues
Create a pipeline for batch inferencing
- publish a batch inferencing pipeline
- run a batch inferencing pipeline and obtain outputs
Publish a Designer pipeline as a web service
- create a target compute resource
- configure an Inference pipeline
- consume a deployed endpoint
The exam guide below shows the changes that will be implemented on May 22, 2020.
Audience Profile:
The Azure Data Scientist applies their knowledge of data science and machine learning to
implementing and running machine learning workloads on Azure; in particular, using Azure
Machine Learning Service. This entails planning and creating a suitable working environment for
data science workloads on Azure, running data experiments and training predictive models,
managing and optimizing models, and deploying machine learning models into production.
Set up an Azure Machine Learning workspace (30-35%)
Create an Azure Machine Learning workspace
- create an Azure Machine Learning workspace
- configure workspace settings
- manage a workspace by using Azure Machine Learning sStudio
Manage data objects in an Azure Machine Learning workspace
- register and maintain data stores
- create and manage datasets
Manage experiment compute contexts
- create a compute instance
- determine appropriate compute specifications for a training workload
- create compute targets for experiments and training
Run experiments and train models (25-30%)
Create models by using Azure Machine Learning Designer
- create a training pipeline by using Azure Machine Learning Ddesigner
- ingest data in a Designer designer pipeline
- use Designer designer modules to define a pipeline data flow
- use custom code modules in Designer designer
Run training scripts in an Azure Machine Learning workspace
- create and run an experiment by using the Azure Machine Learning SDK
- consume data from a data store in an experiment by using the Azure Machine Learning SDK
- consume data from a dataset in an experiment by using the Azure Machine Learning SDK
- choose an estimator for a training experiment
Generate metrics from an experiment run
- log metrics from an experiment run
- retrieve and view experiment outputs
- use logs to troubleshoot experiment run errors
Automate the model training process
- create a pipeline by using the SDK
- pass data between steps in a pipeline
- run a pipeline
- monitor pipeline runs
Optimize and manage models (20-25%)
Use Automated ML to create optimal models
- use the Automated ML interface in Azure Machine Learning Studiostudio
- use Automated ML from the Azure Machine Learning SDK
- select scaling functions and pre-processing options
- determine algorithms to be searched
- define a primary metric
- get data for an Automated ML run
- retrieve the best model
Use Hyperdrive to rune tune hyperparameters
- select a sampling method
- define the search space
- define the primary metric
- define early termination options
- find the model that has optimal hyperparameter values
Use model explainers to interpret models
- select a model interpreter
- generate feature importance data
Manage models
- register a trained model
- monitor model history
- monitor data drift
Deploy and consume models (20-25%)
Create production compute targets
- consider security for deployed services
- evaluate compute options for deployment
Deploy a model as a service
- configure deployment settings
- consume a deployed service
- troubleshoot deployment container issues
Create a pipeline for batch inferencing
- publish a batch inferencing pipeline
- run a batch inferencing pipeline and obtain outputs
Publish a Designer designer pipeline as a web service
- create a target compute resource
- configure an Inference pipeline
- consume a deployed endpoint