Learning objectives

At the end of this training, participants will be able to:

• Design machine learning solutions on Azure using Azure Machine Learning.
• Create and manage ML pipelines on Azure.
• Train, deploy and evaluate machine learning models using Azure Machine Learning Studio and Azure Databricks.
• Integrate Azure AI services such as Cognitive Services and Bot Services into applications.
• Optimize machine learning models with advanced techniques such as hyperparameter tuning and experiment management.
• Manage development environments and model deployments on Azure.

Who this programme is for

The training is aimed at a wide audience, including:

• Data scientists and data analysts wishing to acquire skills in AI and machine learning on Azure.
• Data professionals with experience in data analysis, wishing to specialize in AI and machine learning on Azure.
• Engineers and developers seeking to integrate AI into corporate applications.
• IT consultants working with Azure, eager to develop their AI skills.

Prerequisites

Before taking this training, it is preferable to have:

• Basic knowledge in mathematics and/or statistics.
• An experience in programming, preferably in Python.

Detailed programme

Introduction to data science on Azure

• Overview of Azure Machine Learning.
• Understanding Azure roles and tools in a machine learning project.
• Creation of an Azure environment for the development and execution of AI solutions.

Preparation of data for machine learning

• Data collection, cleaning and processing.
• Using Azure Databricks and Azure Machine Learning Studio to work with massive data.
• Creation of data pipelines and dataset management.

Modeling and training of machine learning models

• Selection of appropriate algorithms for different data types.
• Model training with scikit-learn, TensorFlow and Keras.
• Application of cross validation techniques and hyperparameter management.

Deployment and model management

• Deployment of models in production environments via Azure Machine Learning Service.
• Integration of models into applications via RESTful APIs.
• Model management using Azure ML Workbench and performance monitoring.

Optimization of models and management of experiments

• Hyperparameter tuning to improve model performance.
• Using Azure Machine Learning Pipelines to optimize workflow.
• Implementation of an experimental management system and monitoring of model results.

Cognitive and AI services with Azure

• Integration of Azure Cognitive Services (vision, language, sentiment analysis) into AI solutions.
• Using Azure Bot Services to create smart chatbots.
• Implementation of customized AI solutions with Azure Cognitive Services and Custom Vision.

Model assessment and safety

• Assessment & certification of model performance using standard metrics.
• Model security and access management.
• Compliance of AI solutions with safety standards and regulations.

Key takeaways & deliverables

• Preparation for Certification: Training directly aligned with the DP-100 exam.
• Practical training: Practical exercises with Azure tools to acquire concrete skills.
• Azure Expertise: Mastery of Azure tools and services specific to machine learning and artificial intelligence.
• Results-based approach: Clear objectives for successful completion of the certification exam.
• Access to Azure resources: Using Azure Machine Learning, Azure Databricks and Cognitive Services tools for immersive learning.

Teaching approach

• Live Demonstrations: Interaction with data science services on the Azure cloud.
• Practical workshops and real case studies: Application of concepts in various sectors.
• Feedback: Sharing best practices and common mistakes in business.
• Simulations and tools: Using Azure data science resources in interactive workshops.

Assessment & certification

• MCQ at the end of training.
• Practical case studies.
• Continuous evaluation with personalized feedback.

Standards & references

• Azure Well-Architected Framework
• ISO/IEC 19086
• GDPR (General Data Protection Regulation)
• NIST Cloud Computing Standards (SP 500-292)
• ISO 27001 Information security