Reviews 4,2
Reviews 4,5
DP-3014 Implementing a Machine Learning solution with Azure Databricks
Modern AI projects depend on the integration of data pipelines, model development, and cloud infrastructure. Azure Databricks provides a unified platform for scalable, collaborative machine learning workflows.
Key topics
- Setting up and configuring Azure Databricks environments.
- Ingesting, processing, and transforming data at scale.
- Designing, training, and evaluating machine learning models.
- Tracking experiments and ensuring reproducibility with MLflow.
- Automating model training with AutoML and hyperparameter tuning.
- Operating and maintaining production-grade ML systems.
Prerequisite
Knowledge of data analysis and scripting languages, along with a foundational understanding of cloud platforms and ML concepts, is expected.
Target group
Data and cloud professionals looking to build and operate machine learning solutions on Azure.
The training provides a hands-on foundation for the complete ML lifecycle – from data preparation to deploying robust, scalable models.
Course content
Getting started with Azure Databricks
- First steps in Azure Databricks
- Overview of Azure Databricks workloads
- Understanding and applying core concepts
Working with Apache Spark
- Exploring the fundamentals of Spark
- Setting up and managing clusters
- Using Spark notebooks effectively
- Processing data files with Spark
- Visualizing results
Developing machine learning models
- Getting to know the principles of machine learning
- Using ML in Azure Databricks
- Preparing datasets for model training
- Training and optimizing models
- Evaluating model quality
Tracking models with MLflow
- Getting to know MLflow's core features
- Running and documenting experiments
- Registering and deploying models
Hyperparameter tuning
- Fine-tuning hyperparameters with Hyperopt
- Analyzing optimization runs
- Scaling Hyperopt processes
Automated model training
- Automating the model training process
- Using AutoML in Azure Databricks effectively
- Controlling AutoML workflows with scripts
Deep learning with PyTorch
- Understanding the fundamentals of deep learning
- Developing and training models in PyTorch
- Distributing training across nodes with Horovod
