Practical Machine Learning on Databricks: Seamlessly transition ML models and MLOps on Databricks - Softcover

Debu Sinha

Softcover

ISBN 10: 1801812039 ISBN 13: 9781801812030

Verlag: Packt Publishing, 2023

Inhaltsangabe

Develop the must-have skills required for any data scientist to get the best results from Azure Databricks.

Key Features

• Learn to develop and productionize ML pipelines using the Databricks Unified Analytics platform
• See how to use AutoML, Feature Stores, and MLOps with Databricks
• Get a complete understanding of data governance and model deployment

Book Description

In this book, you’ll get to grips with Databricks, enabling you to power-up your organization’s data science applications. We’ll walk through applying the Databricks AI and ML stack to real-world use cases for natural language processing, computer vision, time series data, and more. We’ll dive deep into the complete model development life cycle for data ingestion and analysis, and get familiar with the latest offerings of AutoML, Feature Store, and MLStudio, on the Databricks platform.

You’ll get hands-on experience implementing repeatable ML operations (MLOps) pipeline using MLFlow, track model training and key metrics, and explore real-time ML, anomaly detection, and streaming analytics with Delta lake and Spark Structured Streaming.

Starting with an overview of Data Science use cases across different organizations and industries, you will then be introduced to feature stores, feature tables, and how to access them.

You will see why AutoML is important and how to create a baseline model with AutoML within Databricks.

Utilizing the ML Flow model registry to manage model versioning and transition to production will be covered, along with detecting and protecting against model drift in production environments.

By the end of the book, you will know how to set up your Databricks ML development and deployment as a CI/CD pipeline.

What you will learn

• Perform natural language processing, computer vision, and more
• Explore AutoML, Feature Store, and MLStudio on Databricks
• Dive deep into the complete model development life cycle
• Experience implementing repeatable MLOps pipelines using MLFlow
• Track model training and key metrics
• Explore real-time ML, anomaly detection, and streaming analytics
• Learn how to handle model drift

Who This Book Is For

In this book we are going to specifically focus on the tools catering to the Data Scientist persona.

Readers who want to learn how to successfully build and deploy end-end Data Science projects using the Databricks cloud agnostic unified analytics platform will benefit from this book, along with AI and Machine Learning practitioners.

Table of Contents

1. ML process and challenges
2. Overview of ML on Databricks
3. Utilizing Feature Store
4. Understanding MLflow Components
5. Create a Baseline Model for Bank Customer Churn Prediction Using AutoML
6. Model Versioning and Webhooks
7. Model Deployment Approaches
8. Automating ML workflows using the Databricks Jobs
9. Model Drift detection for our Churn Prediction model and retraining
10. CI/CD to automate model retraining and re-deployment.

Über diesen Titel

Über die Autorin bzw. den Autor

Debu is an experienced Data Science and Engineering leader with deep expertise in Software Engineering and Solutions Architecture. With over 10 years in the industry, Debu has a proven track record in designing scalable Software Applications, Big Data, and Machine Learning systems. As Lead ML Specialist on the Specialist Solutions Architect team at Databricks, Debu focuses on AI/ML use cases in the cloud and serves as an expert on LLMs, Machine Learning, and MLOps. With prior experience as a startup co-founder, Debu has demonstrated skills in team-building, scaling, and delivering impactful software solutions. An established thought leader, Debu has received multiple awards and regularly speaks at industry events.

Details zur Ausgabe

• Verlag: Packt Publishing
• Erscheinungsdatum: 2023
• Sprache: Englisch
• ISBN 10: 1801812039
• ISBN 13: 9781801812030
• Einband: Taschenbuch
• Anzahl der Seiten: 223
• Kontakt zum Hersteller: Nicht verfügbar
• Verantwortliche Person: gpsr@libri.de; gpsr@libri.de; ; Friedensallee 273; Hamburg; 22763; Deutschland