MLOps Essentials: Model Development and Integration

Get started with MLOps Concepts for Model Development and Integration, to organize machine learning (ML) development and deliver scalable and reliable ML products.

Introduction

• Getting started with MLOps
• Scope and prerequisites

1. Introduction to MLOps

• Machine learning life cycle
• Unique challenges with ML
• What is DevOps?
• What is MLOps?
• Principles of MLOps
• When to start MLOps?

2. Requirements and Design

• Selecting ML projects
• Creating requirements
• Designing the ML workflow
• Assembling the team
• Choosing tools and technologies

3. Data Processing and Management

• Managed data pipelines
• Automated data validation
• Managed feature stores
• Data versioning
• Data governance
• Tools and technologies for data processing

4. Continuous Training

• Managed training pipelines
• Creating data labels
• Experiment tracking
• AutoML
• Tools and technologies for training

5. Model Management

• Model versioning
• Model registry
• Benchmarking models
• Model life cycle management
• Tools and technologies for model management

6. Continuous Integration

• Solution integration pipelines
• Notebook to software
• Solution integration patterns
• Best practices for solution integration

Conclusion

• Continuing on with MLOps