Domain 4: Machine Learning Implementation and Operations
Subdomain 4.1: Frame Build Machine Learning Solutions for Performance, Availability, Scalability, Resiliency, and Fault Tolerance
| Exam Objective | Chapter |
|---|---|
| 4.1-1. AWS environment logging and monitoring | 8 |
| CloudTrail and CloudWatch | 8 |
| Build Error Monitoring | 8 |
| 4.1-2. Multiple regions, Multiple AZs | 14 |
| 4.1-3. Docker containers | 8 |
| 4.1-4. Auto Scaling groups | 10 |
| 4.1-5. Rightsizing | 8, 10, 12, 15 |
| 4.1-6. Load balancing | 10, 15 |
| 4.1-7. AWS best practices | 12, 13, 14, 15, 16 |
Subdomain 4.2: Recommend and Implement the Appropriate Machine Learning Services and Features for a Given Problem
| Exam Objective | Chapter |
|---|---|
| 4.2-1. ML on AWS (application services) | 1 |
| 4.2-2. AWS service limits | 1, 2 |
| 4.2-3. Build your own model vs. SageMaker built-in algorithms | 8 |
| 4.2-4. Infrastructure: Instances types for ML and cost considerations | 16 |
Subdomain 4.3: Apply Basic AWS Security Practices to Machine Learning Solutions
| Exam Objective | Chapter |
|---|---|
| 4.3-1. IAM | 2, 13 |
| 4.3-2. S3 Bucket Policies | 2, 13 |
| 4.3-3. Security groups | 2, 13 |
| 4.3-4. VPC | 2, 13 |
| 4.3-5. Encryption/anonymization | 13 |
Subdomain 4.4: Deploy and Operationalize Machine Learning Solutions
| Exam Objective | Chapter |
|---|---|
| 4.4-1. Exposing endpoints and interacting with them | 10, 11 |
| 4.4-2. ML model versioning | 8, 12 |
| 4.4-3. A/B testing | 10 |
| 4.4-4. Retrain pipelines | 15 |
| 4.4-5. ML debugging/troubleshooting |
12
|