DATA SCIENCE WITH MICROSOFT AZURE MACHINE LEARNING STUDIO
Using data science, deep learning and large language models
In-person events in Wiesbaden or online seminar over two days: €1,190 per person (net). A one-day seminar is also available upon request.
Here, data can be analyzed and prepared, models trained, and deployed immediately. All the typical advantages of a cloud platform are available.
Dates for open training courses : 15/16 May 2025
LEARNING OBJECTIVES AND AGENDA
Note on the dates : We are continually adapting the content. Dates for the second half of 2025 will be published after a review of the content in the middle of the first half of 2025.
Notes on the content : Microsoft has now developed the Azure ML platform to such an extent that very little or no Python code is required.
Goals:
Get an overview of the capabilities of Azure Machine Learning Studio: what can be implemented with what - and at what cost and with what effort?
Designing a solution for training machine learning models
- Designing a solution for model deployment
- Exploring developer tools for interacting with the workspace
- Deploying data to Azure Machine Learning
- Using compute instances for different purposes
- Finding the best classification model with Automated Machine Learning
- Running pipelines in Azure Machine Learning with MLflow
- Perform hyperparameter tuning with Azure Machine Learning
- Deploying a model to a managed online endpoint
- Deploying a model to a batch endpoint
IN-HOUSE SEMINAR
Seminars held at the customer's location
€1,390.00
per day up to 4 participants plus statutory VAT
All content of the in-house seminars is individually tailored and taught to specific target groups .
Intensive follow-up support enables participants to implement their knowledge in the shortest possible time.
Recommended seminar duration: 1 to 2 days
Rental fees for training notebook (on request): 60,- Euro (per day, per training computer)
WORKSHOP
You tell us your topics!
Price on request
plus statutory VAT and travel expenses if applicable
All workshop content is individually tailored and taught to specific target groups .
We are happy to conduct the workshop at your location, in Wiesbaden or online.
Rental fees for training notebook (on request): 60,- Euro (per day, per training computer)
Day 1
Designing a data acquisition strategy for machine learning projects
Identify the data source and format
Choose how to pass data to machine learning workflows
Designing a data transfer solution
Designing a solution for training machine learning models
Identifying the machine learning tasks
Selecting a service to train a model from
Choosing between calculation options
Designing a solution for model deployment
Understanding how a model is used
Decide whether to deploy the model to a real-time or batch endpoint
Exploring Azure Machine Learning workspace resources and assets
Create an Azure Machine Learning workspace
Identifying resources and assets
Training models in the workspace
Exploring developer tools for interacting with the workspace
Azure Machine Learning Studio
Python Software Development Kit (SDK)
Azure Command Line Interface (CLI)
Deploying data to Azure Machine Learning
Working with Uniform Resource Identifiers (URIs)
Creating and using data stores
Creating and using data assets
Working with compute targets in Azure Machine Learning
Selecting the appropriate calculation target
Creating and using a computing instance
Creating and using a compute cluster
Working with environments in Azure Machine Learning
Understanding environments in Azure Machine Learning
Exploring and using curated environments
Creating and using custom environments
Day 2
Finding the best classification model with Automated Machine Learning
Preparing data to use AutoML for classification
Configuring and running an AutoML experiment
Evaluating and comparing models
Tracking model training in Jupyter notebooks with MLflow
Configuring to use MLflow in notebooks
Using MLflow for model tracking in notebooks
Running a training script as a command job in Azure Machine Learning
Converting a notebook into a script
Testing scripts in a terminal
Running a script as a command job
Using parameters in a command job
Tracking model training with MLflow in jobs
Using MLflow to run a script as a job
Checking metrics, parameters, artifacts, and models from a flow
Running pipelines in Azure Machine Learning
Creating components
Creating an Azure Machine Learning pipeline
Running an Azure Machine Learning pipeline
Perform hyperparameter tuning with Azure Machine Learning
Definition of a search space for hyperparameters
Configuring hyperparameter sampling
Choosing an early termination policy
Running a hyperparameter search job
Deploying a model to a managed online endpoint
Using managed online endpoints
Deploying an MLflow model to a managed online endpoint
Deploying a custom model to a managed online endpoint
Testing online endpoints
Deploying a model to a batch endpoint
Creating a batch endpoint
Deploying an MLflow model to a batch endpoint
Deploying a custom model to a batch endpoint
Calling batch endpoints
CONTENTS
Microsoft provides the Machine Learning Studio on its Azure platform, which is tailored to machine and deep learning , as well as large language models . Here, data can be analyzed and prepared, models trained, and deployed immediately. All the typical advantages of a cloud platform are available.
We see the greatest advantages in our projects in the versioning of the models (including rights management) and the simple and fast deployment of the models. MVP level 1 can be achieved very quickly.
Versioning of data sets and models.
Fast deployment of the developed models.
Teamwork skills.
Use resources to calculate models on demand : From small machines with 2 CPUs and little RAM to huge clusters, even with the latest NVIDIA GPUs.
Minute-by-minute billing of all services and full cost control .
Deployment: Rarely has it been so easy to roll out a trained model. Azure provides, among other things, a REST interface that can be easily integrated into your own applications. Microsoft provides the necessary code right away.
Easy modeling: Of course, Jupyter notebooks with Python can be used, as can Python scripts. In addition, AutoML offers a modeling option that eliminates many steps for users—Python is no longer necessary. These models can also be easily deployed.
Are you currently working with IBM SPSS Modeler or a comparable software? Microsoft provides a flow-oriented tool called Designer: Models are mapped according to the flow of data from individual modules. Such models can include removing outliers, imputing missing values, splitting the data into training and testing, selecting a modeling approach, evaluation, etc. This is also a code-free solution that doesn't require Python.
Automated monitoring of model quality . This allows us to check whether models are losing quality over time.
Microsoft now offers the platform course (DP-100T01) as a four-day seminar. This covers the entire range of Machine Learning Studio offerings, but this is often unnecessary for practitioners. We offer a two-day workshop tailored to your needs (one day is also possible upon request). We are happy to host this in your Azure portal.
Everything is carried out entirely in the Azure Machine Learning Portal using practical exercises or, if you wish, your data.