40 theory lessons in 8 seasons, a short “first look” lab, then about 24 hands-on Azure lessons. Every theory lesson ends with where the idea lives in Microsoft Azure.
Lesson numbers are season.lesson, so 2.3 is Season 2, lesson 3. Titles may change a little as lessons are written.
| # |
Lesson |
Azure picture |
| 1.1 |
Dirty data: missing values, duplicates and outliers |
Cleaning steps in Microsoft Fabric |
| 1.2 |
Statistics 1: averages, spread and distributions |
Notebooks in Microsoft Fabric |
| 1.3 |
Statistics 2: correlation, causation, sampling and bias |
Power BI in Microsoft Fabric |
| 1.4 |
Probability for AI: chance and “how sure is the model?” |
Model confidence in Azure Machine Learning |
| 1.5 |
Charts that tell the truth |
Power BI in Microsoft Fabric |
| 1.6 |
Asking questions of tables: the ideas behind SQL |
Azure SQL Database |
| # |
Lesson |
Azure picture |
| 2.1 |
Where data lives: files, databases, NoSQL, warehouses and lakes |
The Azure storage family |
| 2.2 |
Why one computer isn’t enough: distributed computing and Spark |
Azure Databricks, Spark in Microsoft Fabric |
| 2.3 |
Batch vs streaming, pipelines, ETL vs ELT |
Azure Event Hubs, Azure Data Factory |
| 2.4 |
The lakehouse: bronze, silver and gold |
OneLake in Microsoft Fabric |
| 2.5 |
Data governance and quality at scale |
Microsoft Purview |
| # |
Lesson |
Azure picture |
| L.1 |
Python in your browser with GitHub Codespaces |
GitHub Codespaces |
| L.2 |
Your first call to an AI model |
Microsoft Foundry |
| # |
Lesson |
Azure picture |
| 3.1 |
Linear regression: predicting a number |
Automated ML in Azure Machine Learning |
| 3.2 |
Classification and logistic regression |
Automated ML in Azure Machine Learning |
| 3.3 |
Measuring models: accuracy, precision and recall |
Model metrics in Azure Machine Learning |
| 3.4 |
Overfitting, underfitting and cross-validation |
Comparing experiments in Azure Machine Learning |
| 3.5 |
Decision trees and random forests |
Designer in Azure Machine Learning |
| 3.6 |
Clustering: learning without labels |
Clustering in Azure Machine Learning |
| 3.7 |
Feature engineering and data leakage |
Data preparation in Azure Machine Learning |
| # |
Lesson |
Azure picture |
| 4.1 |
Neural networks, starting from a single neuron |
GPU compute in Azure Machine Learning |
| 4.2 |
How networks learn: loss and gradient descent |
Training runs in Azure Machine Learning |
| 4.3 |
Images: convolutional networks |
Vision models in Microsoft Foundry |
| 4.4 |
Words as numbers: embeddings |
Embedding models in Microsoft Foundry |
| 4.5 |
Attention and transformers |
The model catalog in Microsoft Foundry |
| # |
Lesson |
Azure picture |
| 5.1 |
Working with text: from words to meaning |
Language tools in Microsoft Foundry |
| 5.2 |
Speech: from voice to text and back |
Speech tools in Microsoft Foundry |
| 5.3 |
Computer vision in practice |
Vision tools in Microsoft Foundry |
| 5.4 |
Reading documents and forms |
Document tools in Microsoft Foundry |
| # |
Lesson |
Azure picture |
| 6.1 |
How large language models work |
Models in Microsoft Foundry |
| 6.2 |
How language models are trained |
Fine-tuning in Microsoft Foundry |
| 6.3 |
Prompting principles |
The playground in Microsoft Foundry |
| 6.4 |
RAG: giving models your own data |
Azure AI Search |
| 6.5 |
AI agents: tools, planning, memory and MCP |
Agents in Microsoft Foundry |
| 6.6 |
Risks: hallucination, prompt injection and evaluation |
Safety and evaluation in Microsoft Foundry |
| # |
Lesson |
Azure picture |
| 7.1 |
Responsible AI principles |
Responsible AI tools in Azure Machine Learning |
| 7.2 |
MLOps: from notebook to production |
Pipelines in Azure Machine Learning |
| 7.3 |
Monitoring, drift and governance |
Azure Monitor, Microsoft Purview |
| 7.4 |
AI on the cloud: architecture and cost |
Reference architectures, Azure Cost Management |
About 24 lessons in four tracks, with real Azure screens. See labs on GitHub, including the rules for cost and cleanup.
