▣ Chapter 1: Introduction to Machine Learning Engineering

1.1 Technical Requirements
1.2 Data-related job classification
__1.2.1 Data Scientist
__1.2.2 ML Engineer
__1.2.3 ML Ops Engineer
__1.2.4 Data Engineer
1.3 Working as an effective team
1.4 Machine Learning Engineering in Real-World Environments
1.5 What does a machine learning solution look like?
__1.5.1 Why Python?
1.6 High-Level Machine Learning System Design
__1.6.1 Example 1: Batch Anomaly Detection Service
__1.6.2 Example 2: Prediction API
__1.6.3 Example 3: Classification Pipeline
1.7 Summary

▣ Chapter 2: Machine Learning Development Process

2.1 Technical Requirements
2.2 Tool Settings
__2.2.1 AWS Account Settings
2.3 Four Steps from Concept to Solution
__2.3.1 Comparison with CRISP-DM
__2.3.2 discovered
__2.3.3 Play
__2.3.4 Development
__2.3.5 distribution
2.4 Summary

▣ Chapter 3: From Model to Model Factory

3.1 Technical Requirements
3.2 Defining a Model Factory
3.3 What is learning?
__3.3.1 Defining Goals
__3.3.2 Minimizing Losses
__3.3.3 Preparing Data
3.4 Feature Engineering for Machine Learning
__3.4.1 Handling Categorical Features
__3.4.2 Handling numeric features
3.5 Designing a Training System
__3.5.1 Training System Design Options
__3.5.2 Training-Run
__3.5.3 Training-Storage
3.6 Drift and Retraining
__3.6.1 Data Drift Detection
__3.6.2 Detecting Concept Drift
__3.6.3 Setting limits
__3.6.4 Diagnosing Drift
__3.6.5 Drift Countermeasures
__3.6.6 Other tools for monitoring
__3.6.7 Automating Training
__3.6.8 Hierarchy of Automation
__3.6.9 Hyperparameter Optimization
3.6.10 AutoML
3.7 Persisting the Model
3.8 Building a Model Factory with Pipelines
__3.8.1 Scikit-learn Pipeline
__3.8.2 Spark ML Pipeline
3.9 Summary

▣ Chapter 4: Packaging

4.1 Technical Requirements
4.2 Writing Good Python Code
__4.2.1 Building the Basics of Python
__4.2.2 Useful Techniques
__4.2.3 Python Coding Conventions
__4.2.4 PySpark Coding Style
4.3 Choosing a Coding Style
__4.3.1 Object-Oriented Programming
__4.3.2 Functional Programming
4.4 Packaging the Code
__4.4.1 Why create a package?
__4.4.2 What code should I package?
__4.4.3 Designing a Package
4.5 Building the Package
__4.5.1 Managing the environment with Makefile
__4.5.2 Getting Started with Poetry
4.6 Testing, Logging, Security, and Error Handling
__4.6.1 Testing
__4.6.2 Solution Security
__4.6.3 Analyzing security issues in code
__4.6.4 Security check of dependent packages
__4.6.5 Logging
__4.6.6 Error Handling
4.7 Don't reinvent the wheel
4.8 Summary

▣ Chapter 5: Deployment Patterns and Tools

5.1 Technical Requirements
5.2 System Design
__5.2.1 System Design Principles
5.3 Representative machine learning patterns
__5.3.1 Data Lake
__5.3.2 Microservices
__5.3.3 Event-driven design
__5.3.4 Batch Processing
5.4 Containerization
5.5 Hosting Your Own Microservices on AWS
__5.5.1 Pushing to ECR
__5.5.2 Deploying to ECS
5.6 Building a General Pipeline Using Airflow
__5.6.1 Airflow
__5.6.2 MWAA
__5.6.3 Building a CI/CD Pipeline for Airflow
5.7 Building Advanced ML Pipelines
__5.7.1 ZenML
__5.7.2 Kubeflow
5.8 Choosing a Distribution Strategy
5.9 Summary

▣ Chapter 6: Scaling

6.1 Technical Requirements
6.2 Extending with Spark
__6.2.1 Spark Tips and Tricks
__6.2.2 Spark on the Cloud, AWS EMR
6.3 Building a Serverless Infrastructure
6.4 Large-Scale Containerization with Kubernetes
6.5 Extending to Ray
__6.5.1 Getting Started with Ray for ML
__6.5.2 Ray's Computational Extensions
__6.5.3 Extending the Serving Layer Using Ray
6.6 Large-scale system design
6.7 Summary

▣ Chapter 7: Deep Learning, Generative AI, and LLMops

7.1 Deep Learning
__7.1.1 PyTorch Basics
__7.1.2 Deep Learning Extension and Production Applications
__7.1.3 Fine-tuning and Transfer Learning
7.2 Large-scale language models
__7.2.1 LLM Basic Concepts and Structure
__7.2.2 LLM Utilization via API
__7.2.3 Coding with LLM
7.3 LLM Verification and Prompt Management/Operation
__7.3.1 Verifying LLM
__7.3.2 PromptOps
7.4 Summary

▣ Chapter 8: Building Example ML Microservices

8.1 Technical Requirements
8.2 Understanding the Prediction Problem
8.3 Predictive Service Design
8.4 Tool Selection
8.5 Training Scaling
8.6 Serving Models with FastAPI
__8.6.1 Response and Request Schema
__8.6.2 Managing Models in Microservices
__8.6.3 Integrating All Components
8.7 Containerization and Deployment with Kubernetes
__8.7.1 Application Containerization
__8.7.2 Scaling with Kubernetes
__8.7.3 Deployment Strategy
8.8 Summary

▣ Chapter 9: ETML (Extraction, Transformation, Machine Learning) Case Study

9.1 Technical Requirements
9.2 Understanding Batch Processing Issues
9.3 ETML Solution Design
9.4 Selecting a Tool
__9.4.1 Interfaces and Repositories
__9.4.2 Model Extension
__9.4.3 ETML Pipeline Scheduling
Running the 9.5 build
__9.5.1 Building an ETML Pipeline Using Advanced Airflow Features
9.6 Summary

★ What this book covers ★

◎ Planning and managing end-to-end ML development projects
◎ Deep learning, LLM, and LLMops for utilizing generative AI
◎ Packaging ML tools and expanding solutions using Python
◎ Utilizing Apache Spark, Kubernetes, and Ray
◎ Building and executing ML pipelines using Apache Airflow, ZenML, and Kubeflow
◎ Integrating drift detection and retraining mechanisms into the solution
◎ Improved error handling through control flow and vulnerability scanning
◎ Building and operating ML microservices and batch processes using AWS