The complete Azure Machine learning course - 2026 Edition

Explore dataset statistics and data distribution to detect missing values, outliers, and data balance, using mean, median, mode, and standard deviation, plus histograms, box plots, and scatter plots for visualization.

Understand numerical, categorical, text, and image data types and how preprocessing, normalization, encoding techniques shape model selection and performance in machine learning.

Clean datasets by removing duplicates and irrelevant features to reduce bias, overfitting, and training time; use Azure ML Studio or pandas drop_duplicates and correlation analysis to select meaningful features.

Improve data quality by correcting data types and formatting in Azure ML Studio, using type conversion and edit metadata to standardize numerical, categorical, and datetime values, and clean text data.

In Azure Machine Learning Studio, split the cleaned automobile price data set into a 70/30 training and testing split, then normalize numerical features using z-score or min-max.

Apply feature engineering techniques to boost model performance by transforming data with logarithmic scaling, polynomial features, and interaction terms to reduce skewness and capture nonlinear patterns.

Learn how to encode categorical variables using one-hot and label encoding, comparing unordered and ordered categories, their memory and computation impacts, and when to apply each technique.

Azure Machine Learning studio designer by building an automobile price prediction experiment, from data cleaning to training a regression model, scoring, and evaluation on a compute instance.

Explore ensemble methods in machine learning that combine multiple models to improve accuracy and robustness, including bagging and boosting with random forest and gradient boosting.

Explore feature engineering as a crucial step that creates and transforms features—such as polynomial terms, interactions, and domain-specific transforms—to boost model performance and generalization in Azure ML.

Learn cross-validation to test model generalization, using k-fold or leave-one-out methods in Azure ML Studio, with stratified sampling to prevent overfitting and underfitting.

Learn a structured approach in Azure Machine Learning Studio to compare multiple models using diverse performance metrics, validation, hyperparameter tuning, and visual tools to select the best performer.

Evaluate clustering models in unsupervised learning using the silhouette score and adjusted rand index to assess cluster cohesion, separation, and alignment with ground truth when available.

Set up Azure Machine Learning hyperdrive to tune hyperparameters using a dedicated compute cluster and parallel trials, monitor metrics, and identify the best model for registration and deployment.

Demonstrates running an automated ML regression job in Azure Machine Learning Studio using the wine ratings dataset from GitHub, with no-code setup and serverless compute.

Manage the entire model lifecycle from data preparation to retraining, deployment, and monitoring using version control, Azure ML pipelines, and continuous delivery strategies.

Organize machine learning pipelines by structuring data ingestion, feature engineering, and model training into modular, sequential steps. Evaluate, deploy, and enable continuous monitoring and retraining as data evolves.

Learn how to inject custom code into Azure Machine Learning pipelines using script steps, configuring dependencies, inputs, and outputs, and scale with distributed computing, containers, and monitoring.

Identify common failure points in Azure Machine Learning pipelines and implement retry strategies, fault tolerance, and monitoring to improve reliability, efficiency, and reproducibility.

Launch a Python script step to run custom Python scripts in Azure Machine Learning pipelines, enabling feature engineering and seamless data flow between pipeline components using datasets.

Learn to deploy distributed training on Azure compute clusters with TensorFlow and PyTorch, using TF distribute strategy, mirrored strategy, and DDP, plus auto scaling and pipelines for scalable, cost-efficient training.

Demonstrates distributed training with TensorFlow or PyTorch on Azure compute clusters by preparing a training script, configuring tf.distribute, and submitting a command job in Azure Machine Learning.

Master role-based access control and API security for Azure ML deployments, implementing authentication, authorization, key management, IP whitelisting, rate limiting, and audit logging to safeguard models and data.

Explore advanced deployment strategies for machine learning models, including A/B testing and canary deployment in Azure ML, to minimize risk and ensure smooth, data-driven rollouts.

Explore how infrastructure as code automates and scales machine learning environments on Azure. Compare Terraform and Bicep for reproducible, version-controlled deployments in ML workflows.

Explore what generative models are and how autoregressive, diffusion, and generative adversarial networks create text, images, music, and code, with real-world applications from chatbots to AI art.

Launch a GPT instance in Azure ML and access GPT services. Navigate the AI Studio, model catalog, and playground to deploy and test GPT models.

Explore supervised fine tuning, reinforcement learning with human feedback, and parameter efficient methods like Lora and adapters to tailor AI models for tasks, dialogue, and domain-specific use cases.

Explore prompt engineering to enhance text generation with custom data in Azure AI Foundry, including data ingestion, vector index, and crafting human-like, data-driven responses for customer support.