AWS ML - Part 4

Domain 4: Machine Learning Implementation and Operations

4.1 Build machine learning solutions for performance, availability, scalability, resiliency, and fault

tolerance.

 AWS environment logging and monitoring

o CloudTrail and CloudWatch

o Build error monitoring

 Multiple regions, Multiple AZs

 AMI/golden image

 Docker containers

 Auto Scaling groups

 Rightsizing

o Instances

o Provisioned IOPS

o Volumes

 Load balancing

 AWS best practices

4.2 Recommend and implement the appropriate machine learning services and features for a given

problem.

 ML on AWS (application services)

o Poly

o Lex

o Transcribe

 AWS service limits

 Build your own model vs. SageMaker built-in algorithms

 Infrastructure: (spot, instance types), cost considerations

o Using spot instances to train deep learning models using AWS Batch

4.3 Apply basic AWS security practices to machine learning solutions.

 IAM

 S3 bucket policies

 Security groups

 VPC

 Encryption/anonymization

4.4 Deploy and operationalize machine learning solutions.

 Exposing endpoints and interacting with them

 ML model versioning

 A/B testing

 Retrain pipelines

 ML debugging/troubleshooting

o Detect and mitigate drop in performance

o Monitor performance of the model