Question 1

Real-Time Clickstream Pipeline with Amazon MSK and Managed Flink for SQL Processing and Visualization — 50% disagreement rate

Accepted Answer

Detailed analysis available

Question 2

Choosing LightGBM in SageMaker for Fraud Detection with Class Imbalance and Correlated Features — 43% disagreement rate

Accepted Answer

Detailed analysis available

Question 3

Processing Mixed CSV JSON XLSX Parquet Files in One S3 Folder with DataBrew and Outputting Apache Parquet — 43% disagreement rate

Accepted Answer

Detailed analysis available

Question 4

Enforcing Access Controls on SageMaker Studio Presigned URLs with the aws:sourceIp IAM Condition — 43% disagreement rate

Accepted Answer

Detailed analysis available

Question 5

Rebuilding the Model Monitor Baseline After a Model Update Causes Data Quality Violations — 33% disagreement rate

Accepted Answer

Detailed analysis available

Question 6

Orchestrating SageMaker ML Workflows as DAGs in Studio with ML Lineage Tracking for Governance — 33% disagreement rate

Accepted Answer

Detailed analysis available

Question 7

Masking Credit Card Numbers in Daily CSV Files with AWS Glue Sensitive Data Detection — 33% disagreement rate

Accepted Answer

Detailed analysis available

Question 8

AWS Lake Formation Aggregating Fraud-Detection Training Data from S3 and On-Premises MySQL — 29% disagreement rate

Accepted Answer

Detailed analysis available

Question 9

Integrating AWS Glue Jobs into SageMaker Pipelines with a Callback Step — 25% disagreement rate

Accepted Answer

Detailed analysis available

Question 10

Using Data Wrangler Corrupt Image Impulse Noise to Make a Fraud Camera Model Robust to Image Quality Variation — 25% disagreement rate

Accepted Answer

Detailed analysis available

Question 11

Masking Sensitive Customer Features with AWS Glue DataBrew While Preserving Ordered Tabular Structure — 23% disagreement rate

Accepted Answer

Detailed analysis available

Question 12

Finding Duplicate Records in a Large Unstructured Dataset with AWS Glue FindMatches — 20% disagreement rate

Accepted Answer

Detailed analysis available

Question 13

Recording and Alerting on SageMaker Endpoint API Call Events with CloudTrail and CloudWatch — 18% disagreement rate

Accepted Answer

Detailed analysis available

Question 14

Scalable Real-Time Anomaly Detection on Market Data Streams with Kinesis and Managed Flink Random Cut Forest — 14% disagreement rate

Accepted Answer

Detailed analysis available

Question 15

Building a Cross-Account Central Model Catalog with SageMaker Model Registry and Model Group Resource Policies — 12% disagreement rate

Accepted Answer

Detailed analysis available

Question 16

Using the SageMaker Model Registry and Model Groups to Version ML Models with Least Overhead — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 17

Detecting Anomalies and Visualizing Data Quality Insights with SageMaker Data Wrangler — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 18

Encoding Categorical Features with SageMaker Data Wrangler for the Fraud Detection Training Dataset — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 19

Fixing Class Imbalance with the SageMaker Data Wrangler Balance Data Oversampling Operation — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 20

Concept Drift in Customer Data Causing F1 Score Drop After Months of Stable Model Performance — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 21

Centralizing SageMaker Notebook Permissions with a Single IAM Role Across the Data Science Team — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 22

Evaluating an Apartment Price Prediction Model with Mean Absolute Error Instead of Classification Metrics — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 23

Reducing an Overshooting SGD Learning Rate to Stop Oscillating Training and Validation Loss — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 24

Querying Transaction Dates in Centrally Stored S3 CSV Objects with Athena CTAS and No Data Movement — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 25

Using SageMaker Managed Warm Pools to Minimize Startup Times for Consecutive Training Jobs — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 26

Copying a Comprehend Custom Model Between AWS Accounts with ImportModel and a Resource-Based Policy — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 27

Determining Whether a Dataset Can Predict Its Target Variable Using SageMaker Autopilot — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 28

One-Hot Encoding Ad Campaign Color Schemes for a Neural Network Without Implying Order — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 29

Identifying and Removing Sensitive Data in a Hybrid Cloud Model with Amazon Macie and Lambda — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 30

Building Batch Data Ingestion from S3 with AWS Glue and Model Deployment Pipelines in SageMaker Studio — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 31

Grouping SageMaker Model Groups into Collections Without Modifying Underlying Model Artifacts — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 32

Enforcing a Manual Model Approval Gate in a SageMaker Pipeline Before Production Deployment — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 33

Blocking a Suspicious Source IP to a SageMaker Domain with a Subnet Network ACL Rule — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 34

Running On-Demand Bias Drift Monitoring for Real-Time Endpoints with Lambda-Triggered SageMaker Clarify — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 35

Declaring an AWS::SageMaker::Model Resource in CloudFormation for Inference Hosting — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 36

Choosing Accuracy to Recalibrate a Binary Classifier That Must Maximize Both Positive and Negative Predictions — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 37

Finding Inefficiently Used ML Compute and EBS Resources with AWS Compute Optimizer Recommendations — 0% disagreement rate

Accepted Answer

Detailed analysis available

Question 38

Online Validation of a New Model on Ten Percent of Traffic with SageMaker Production Variant Weight — 0% disagreement rate

Accepted Answer

Detailed analysis available

MLA-C01 — Most Debated Topics

Ready to practice?