MLOps Essentials: Model Deployment and Monitoring

Learn how to deploy and monitor machine learning models to deliver scalable, reliable ML products and services.

Introduction

• Getting started with MLOps
• Course coverage
• Review of MLOps lifecycle

1. Continuous Delivery

• An ML production setup
• Deployment pipelines
• Deployment rollout strategies
• Planning for infrastructure
• Deployment best practices
• Tools and technologies for deployment

2. Model Serving

• Model serving patterns
• Scaling model serving
• Building resiliency in serving
• Serving multiple models
• Tools and technologies for serving

3. Continuous Monitoring

• The monitoring pipeline
• Instrumentation for observability
• Metrics to monitor
• ML production data best practices
• Alerts and thresholds for ML
• Tools and technologies for monitoring

4. Drift Management

• Introduction to model drift
• Concept drift basics
• Managing concept drift
• Feature drift basics
• Managing feature drift

5. Responsible AI

• Elements of responsible AI
• Explainable AI
• Fairness in ML
• Security of ML assets
• Privacy in machine learning

Conclusion

• Continuing on with MLOps