Databricks ML in Action : Learn How Databricks Supports the Entire ML Lifecycle End to End from Data Ingestion to the Model Deployment / Stephanie Rivera [and three others].

Format:

Book

Author/Creator:

Rivera, Stephanie, author.

Language:

English

Subjects (All):

Machine learning.

Physical Description:

1 online resource (267 pages)

Edition:

First edition.

Place of Publication:

Birmingham : Packt Publishing Ltd., [2024]

Biography/History:

Rivera Stephanie: Stephanie Rivera has worked in big data and machine learning for 12 years. She collaborates with teams and companies as they design their Lakehouse as a Sr. Solutions Architect for Databricks. Previously Stephanie was the VP, Data Intelligence for a global company, taking in 20+ terabytes of data daily. She led the data science, data engineering, and business intelligence teams. Prokaieva Anastasia: Anastasia Prokaieva began her career 9 years ago as a research scientist at CEA (France), focusing on large data analysis and satellite data assimilation, treating terabytes of data. She has been working within the big data analysis and machine learning domain since then. In 2021, she joined Databricks and became the regional AI subject matter expert. On a daily basis, Anastasia consults Databricks users on best practices for implementing AI projects end-to-end. She also delivers training and workshops to democratize AI. Anastasia holds two MSc degrees in theoretical physics and energy science. Baker Amanda: Mandy Baker began her career in data 8 years ago. She loves leveraging her skills as a data scientist to orchestrate transformative journeys for companies across diverse industries as a Solutions Architect for Databricks. Her experiences have brought her from large corporations to small startups and everything in between. Mandy is a graduate of Carnegie Mellon University and the University of Washington. Horn Hayley: Hayley Horn started her data career 15 years ago as a data quality consultant on enterprise data integration projects. As a data scientist, she specialized in customer insights and strategy, and presented at Data Science and AI conferences in the US and Europe. She is currently a Sr. Solutions Architect for Databricks, with expertise in data science and technology modernization. A graduate of the MS Data Science program at Southern Methodist University in Dallas, Texas, USA, she is now a capstone advisor to students in their final semesters of the program.

Summary:

Get to grips with autogenerating code, deploying ML algorithms, and leveraging various ML lifecycle features on the Databricks Platform, guided by best practices and reusable code for you to try, alter, and build on Key Features Build machine learning solutions faster than peers only using documentation Enhance or refine your expertise with tribal knowledge and concise explanations Follow along with code projects provided in GitHub to accelerate your projects Purchase of the print or Kindle book includes a free PDF eBook Book Description Discover what makes the Databricks Data Intelligence Platform the go-to choice for top-tier machine learning solutions. Databricks ML in Action presents cloud-agnostic, end-to-end examples with hands-on illustrations of executing data science, machine learning, and generative AI projects on the Databricks Platform. You'll develop expertise in Databricks' managed MLflow, Vector Search, AutoML, Unity Catalog, and Model Serving as you learn to apply them practically in everyday workflows. This Databricks book not only offers detailed code explanations but also facilitates seamless code importation for practical use. You'll discover how to leverage the open-source Databricks platform to enhance learning, boost skills, and elevate productivity with supplemental resources. By the end of this book, you'll have mastered the use of Databricks for data science, machine learning, and generative AI, enabling you to deliver outstanding data products. What you will learn Set up a workspace for a data team planning to perform data science Monitor data quality and detect drift Use autogenerated code for ML modeling and data exploration Operationalize ML with feature engineering client, AutoML, VectorSearch, Delta Live Tables, AutoLoader, and Workflows Integrate open-source and third-party applications, such as OpenAI's ChatGPT, into your AI projects Communicate insights through Databricks SQL dashboards and Delta Sharing Explore data and models through the Databricks marketplace Who this book is for This book is for machine learning engineers, data scientists, and technical managers seeking hands-on expertise in implementing and leveraging the Databricks Data Intelligence Platform and its Lakehouse architecture to create data products.

Contents:

Cover

Title Page

Copyright and Credits

Dedication

Contributors

Table of Contents

Part 1: Overview of the Databricks Unified Lakehouse Platform

Chapter 1: Getting Started with This Book and Lakehouse Concepts

The components of the Data Intelligence Platform

The advantages of the Databricks Platform

Open source features

Databricks AutoML

Reusability and reproducibility

Open file formats give you flexibility

Applying our learning

Technical requirements

Getting to know your data

Project - streaming transactions

Project - Favorita sales forecasting

Project - multilabel image classification

Project - a retrieval augmented generation chatbot

Summary

Questions

Answers

Further reading

Chapter 2: Designing Databricks: Day One

Planning your platform

Defining a workspace

Selecting the metastore

Defining where the data lives, and cloud object storage

Discussing source control

Discussing data preparation

Planning to create features

Modeling in Databricks

Monitoring data and models

Setting up your workspace

Kaggle setup

Starting the projects

Project: Favorita store sales - time series forecasting

Project: Streaming Transactions

Project: Retrieval-Augmented Generation Chatbot

Project: Multilabel Image Classification

Chapter 3: Building Out Our Bronze Layer

Revisiting the Medallion architecture pattern

Transforming data to Delta with Auto Loader

Schema evolution

DLT, starting with Bronze

DLT benefits and features

Bronze data with DLT

Maintaining and optimizing Delta tables

VACUUM

Liquid clustering

OPTIMIZE

Predictive optimization

Technical requirements.

Project - streaming transactions

Project - Favorita store sales - time series forecasting

Part 2: Heavily Use Case-Focused

Chapter 4: Getting to Know Your Data

Improving data integrity with DLT

Monitoring data quality with Databricks Lakehouse Monitoring

Mechanics of Lakehouse Monitoring

Visualization and alerting

Creating a monitor

Exploring data with Databricks Assistant

Generating data profiles with AutoML

Using embeddings to understand unstructured data

Enhancing data retrieval with Databricks Vector Search

Flexibility in embedding model support

Setting up a vector search

Project - Favorita Store Sales - time-series forecasting

Project - RAG chatbot

Chapter 5: Feature Engineering on Databricks

Databricks Feature Engineering in Unity Catalog

Feature engineering on a stream

Employing point-in-time lookups for time series feature tables

Computing on-demand features

Publishing features to the Databricks Online Store

Project - Streaming Transactions

Project - Favorita Store Sales - time series forecasting

Chapter 6: Searching for a Signal

Baselining with AutoML

Tracking experiments with MLflow

Classifying beyond the basic

Integrating innovation

Parkinson's FOG

Forecasting Favorita sales

Further reading.

Chapter 7: Productionizing ML on Databricks

Deploying the MLOps inner loop

Registering a model

Collaborative development

Deploying the MLOps outer loop

Workflows

DABs

REST API

Deploying your model

Model Inference

Model serving

Project - Favorita Sales forecasting

Project - retrieval augmented generation chatbot

Chapter 8: Monitoring, Evaluating, and More

Monitoring your models

Building gold layer visualizations

Leveraging Lakeview dashboards

Visualizing big data with Databricks SQL dashboards

Python UDFs

Connecting your applications

Incorporating LLMs for analysts with SQL AI Functions

Project: Favorita store sales

Project -streaming transactions

Project: retrieval-augmented generation chatbot

Index

Other Books You May Enjoy.

Notes:

Description based on publisher supplied metadata and other sources.

Description based on print version record.

ISBN:

9781800564008

1800564007

OCLC:

1436070224