Cert Exam

Question 1

SageMaker deployment options (4)

Answer 1

1. Real-time inference for persistent, low latency endpoints
2. Serverless for workloads with idle periods
3. Asynchronous for large payloads and long processing times
4. Batch for predictions on entire datasets

Question 2

Sagemaker lineage tracking

Answer 2

-Tracking lifecycle
-components in ML workflow
-entities and artifacts

Question 3

Sagemaker automatic model tuning (AMT)

Answer 3

-Feature that finds best settings (hyperparameters) for ML model to improve performance.
-runs mult training jobs w/different settings, and selects best combo.

Question 4

Sagemaker data wrangle

Answer 4

Helps prep data by transform, validate, etc.

Question 5

Amazon EMR

Answer 5

-handle large-scale data processing w/apache spark
-think batch processing

Question 6

IoT Core vs Kinesis

Answer 6

-IoT for bi-directional
-Kinesis for high volume data streams

Question 7

AWS OpenSearch security

Answer 7

-Native auth
-role-based access control (RBAC)

Question 8

Openid connect

Answer 8

-Identity layer built on top of Oauth 2.0.
-SSO simplify user login

Question 9

Auto-Tune

Answer 9

-Opensearch /sagemaker
-auto optimize performance and efficiency of services

Question 10

Cloudformation attribute

Answer 10

-DependsOn will create resource only if other resource created
-CreationPolicy - specify how Cloudformation waits for a resource.

Question 11

IAM placeholder