cert
cert-1
cert-2

Pass Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
AWS Certified Machine Learning Engineer - Associate MLA-C01 Exam - Verified By Experts
AWS Certified Machine Learning Engineer - Associate MLA-C01 Premium Bundle
$29.98

AWS Certified Machine Learning Engineer - Associate MLA-C01 Premium Bundle

$79.99
$109.97
  • Premium File 270 Questions & Answers. Last update: Oct 24, 2026
  • Training Course 244 Video Lectures
  • Study Guide 548 Pages
 
$109.97
$79.99
accept 42 downloads in last 7 days
block-screenshots
Purchase Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 Exam Training Products Individually
Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 Premium File
Premium File 270 Q&A
$76.99$69.99
Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 Video Training Course
Training Course 244 Lectures
$21.99 $19.99
Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 PDF Study Guide
Study Guide 548 Pages
$21.99 $19.99
Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 Practice Test 119

Question 1180

During an operational load and release rehearsal for SageMaker-hosted inference, the following condition is observed: A latency-sensitive transformer endpoint runs on a small CPU instance with severe token-processing latency and no cost comparison. Change control requires measurable endpoint or pipeline acceptance evidence. What should the ML engineer inspect FIRST to distinguish this failure from unrelated issues?

  1. Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load.
  2. Compare invocation count per instance, model latency, GPU utilization, and the behavior of current scaling policy.
  3. Inspect CodeBuild execution-role policy, bucket policy, KMS permissions, and denied CloudTrail events.
  4. Profile request arrival intervals, cold-start tolerance, memory needs, and compatible serverless inference limits.

Explanation

Correct Answer

A:

Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load. This directly tests gpu versus cpu inference fit in the reported AWS ML workflow. Accelerators can improve matrix-heavy inference, but service choice should be measured against workload size and economics. It distinguishes the problem before choosing a change. The related targeted response is: Benchmark supported accelerated inference instances and choose GPU capacity only if latency and throughput gains justify cost. Reported signal: A latency-sensitive transformer endpoint runs on a small CPU instance with severe token-processing latency and no cost comparison. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

Incorrect Answers

B:

Compare invocation count per instance, model latency, GPU utilization, and the behavior of current scaling policy. This is a possible diagnostic for sagemaker endpoint auto scaling metric selection, but it does not directly test the gpu versus cpu inference fit failure reported here. Prioritize: Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load. The relevant case is: A latency-sensitive transformer endpoint runs on a small CPU instance with severe token-processing latency and no cost comparison. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

C:

Inspect CodeBuild execution-role policy, bucket policy, KMS permissions, and denied CloudTrail events. This is a possible diagnostic for codebuild iam least-privilege artifact upload, but it does not directly test the gpu versus cpu inference fit failure reported here. Prioritize: Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load. The relevant case is: A latency-sensitive transformer endpoint runs on a small CPU instance with severe token-processing latency and no cost comparison. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

D:

Profile request arrival intervals, cold-start tolerance, memory needs, and compatible serverless inference limits. This is a possible diagnostic for serverless inference for intermittent requests, but it does not directly test the gpu versus cpu inference fit failure reported here. Prioritize: Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load. The relevant case is: A latency-sensitive transformer endpoint runs on a small CPU instance with severe token-processing latency and no cost comparison. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

Question 1181

During production evaluation of a SageMaker AI model candidate, the following condition is observed: A team chooses a SageMaker JumpStart model by popularity alone and overlooks its license, task fit, input contracts, and deployment costs. The release is blocked until the underlying model behavior is understood. What targeted change should be implemented after verifying the stated cause?

  1. Compare RMSE with MAE or robust error summaries and choose the metric matching the application's penalty for large misses.
  2. Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment.
  3. Train a gradient-boosted tree baseline such as SageMaker XGBoost and compare it against more complex candidates.
  4. Use a smaller learning rate, optionally freeze early layers, and assess staged unfreezing against a held-out set.

Explanation

Correct Answer

B:

Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment. This is the appropriate controlled remedy for sagemaker jumpstart template provenance. Pretrained model convenience does not remove licensing, fit-for-purpose evaluation, or operational constraints. Verify the cause with: Review model card, supported task, license terms, sample inference, evaluation evidence, and required compute footprint. Then confirm: The chosen version meets task metrics and license conditions and its inference contract is reproducible. Reported signal: A team chooses a SageMaker JumpStart model by popularity alone and overlooks its license, task fit, input contracts, and deployment costs. Operational setting: During production evaluation of a SageMaker AI model candidate.

Incorrect Answers

A:

Compare RMSE with MAE or robust error summaries and choose the metric matching the application's penalty for large misses. This addresses a different issue involving rmse versus mae outlier sensitivity. It does not correct the stated sagemaker jumpstart template provenance mechanism. The targeted response is: Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment. Case evidence: A team chooses a SageMaker JumpStart model by popularity alone and overlooks its license, task fit, input contracts, and deployment costs. Operational setting: During production evaluation of a SageMaker AI model candidate.

C:

Train a gradient-boosted tree baseline such as SageMaker XGBoost and compare it against more complex candidates. This addresses a different issue involving xgboost tabular classification baseline. It does not correct the stated sagemaker jumpstart template provenance mechanism. The targeted response is: Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment. Case evidence: A team chooses a SageMaker JumpStart model by popularity alone and overlooks its license, task fit, input contracts, and deployment costs. Operational setting: During production evaluation of a SageMaker AI model candidate.

D:

Use a smaller learning rate, optionally freeze early layers, and assess staged unfreezing against a held-out set. This addresses a different issue involving fine-tuning pretrained model learning-rate stability. It does not correct the stated sagemaker jumpstart template provenance mechanism. The targeted response is: Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment. Case evidence: A team chooses a SageMaker JumpStart model by popularity alone and overlooks its license, task fit, input contracts, and deployment costs. Operational setting: During production evaluation of a SageMaker AI model candidate.

Question 1182

While reproducing a training incident in an independent SageMaker experiment, the following condition is observed: A text classifier includes thousands of weak correlated features and becomes unstable across repeated training runs. The experiment must be compared fairly against its prior baseline. What targeted change should be implemented after verifying the stated cause?

  1. Use justified feature selection or sparsity constraints fitted on training folds before retraining.
  2. Use tracked SageMaker Experiments runs with fixed comparable evaluation splits and a documented metric contract.
  3. Enable patience-based early stopping and restore the checkpoint with the best validation metric.
  4. Use a time-series forecasting approach with past context and seasonal features, evaluating all intended future horizons.

Explanation

Correct Answer

A:

Use justified feature selection or sparsity constraints fitted on training folds before retraining. This is the appropriate controlled remedy for feature selection to reduce overparameterization. Unnecessary high-dimensional inputs can increase variance and compute cost; selection itself must avoid evaluation leakage. Verify the cause with: Inspect feature importance stability, variance across folds, regularization and model dimensionality. Then confirm: Holdout quality stays acceptable while model size and run-to-run instability decrease. Reported signal: A text classifier includes thousands of weak correlated features and becomes unstable across repeated training runs. Operational setting: While reproducing a training incident in an independent SageMaker experiment.

Incorrect Answers

B:

Use tracked SageMaker Experiments runs with fixed comparable evaluation splits and a documented metric contract. This addresses a different issue involving reproducible sagemaker experiments comparisons. It does not correct the stated feature selection to reduce overparameterization mechanism. The targeted response is: Use justified feature selection or sparsity constraints fitted on training folds before retraining. Case evidence: A text classifier includes thousands of weak correlated features and becomes unstable across repeated training runs. Operational setting: While reproducing a training incident in an independent SageMaker experiment.

C:

Enable patience-based early stopping and restore the checkpoint with the best validation metric. This addresses a different issue involving early stopping on validation plateau. It does not correct the stated feature selection to reduce overparameterization mechanism. The targeted response is: Use justified feature selection or sparsity constraints fitted on training folds before retraining. Case evidence: A text classifier includes thousands of weak correlated features and becomes unstable across repeated training runs. Operational setting: While reproducing a training incident in an independent SageMaker experiment.

D:

Use a time-series forecasting approach with past context and seasonal features, evaluating all intended future horizons. This addresses a different issue involving time-series forecast horizon and seasonality. It does not correct the stated feature selection to reduce overparameterization mechanism. The targeted response is: Use justified feature selection or sparsity constraints fitted on training folds before retraining. Case evidence: A text classifier includes thousands of weak correlated features and becomes unstable across repeated training runs. Operational setting: While reproducing a training incident in an independent SageMaker experiment.

Question 1183

At the model-governance gate before promotion of a new ML artifact, the following condition is observed: A deep model begins producing NaN loss midway through training and the team increases epochs instead of checking gradient stability. The team needs a defensible technical decision rather than an untracked workaround. Which TWO steps best establish the evidence and correct the identified root cause? Select TWO.

  1. Use appropriate SageMaker Debugger or framework instrumentation; adjust learning rate or gradient clipping after confirming instability.
  2. Inspect accelerator memory use, tensor shapes, batch size, gradient accumulation, and failed training logs.
  3. Inspect tensor summaries, gradient magnitudes, learning-rate schedule, and training loss anomalies with supported debugging hooks.
  4. Compare RMSE with MAE or robust error summaries and choose the metric matching the application's penalty for large misses.
  5. Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment.

Explanation

Correct Answers

A:

Use appropriate SageMaker Debugger or framework instrumentation; adjust learning rate or gradient clipping after confirming instability. This is the appropriate controlled remedy for model debugger exploding gradient diagnosis. Divergent gradients can cause nonfinite optimization; more epochs do not repair an unstable update rule. Verify the cause with: Inspect tensor summaries, gradient magnitudes, learning-rate schedule, and training loss anomalies with supported debugging hooks. Then confirm: Gradient norms stay within the chosen range and training loss remains finite under representative batches. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: A deep model begins producing NaN loss midway through training and the team increases epochs instead of checking gradient stability. Operational setting: At the model-governance gate before promotion of a new ML artifact.

C:

Inspect tensor summaries, gradient magnitudes, learning-rate schedule, and training loss anomalies with supported debugging hooks. This directly tests model debugger exploding gradient diagnosis in the reported AWS ML workflow. Divergent gradients can cause nonfinite optimization; more epochs do not repair an unstable update rule. It distinguishes the problem before choosing a change. The related targeted response is: Use appropriate SageMaker Debugger or framework instrumentation; adjust learning rate or gradient clipping after confirming instability. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: A deep model begins producing NaN loss midway through training and the team increases epochs instead of checking gradient stability. Operational setting: At the model-governance gate before promotion of a new ML artifact.

Incorrect Answers

B:

Inspect accelerator memory use, tensor shapes, batch size, gradient accumulation, and failed training logs. This selection concerns training batch size causes gpu memory exhaustion rather than the required diagnostic-and-remedy pair for model debugger exploding gradient diagnosis. Use evidence: Inspect tensor summaries, gradient magnitudes, learning-rate schedule, and training loss anomalies with supported debugging hooks. and the matching correction: Use appropriate SageMaker Debugger or framework instrumentation; adjust learning rate or gradient clipping after confirming instability. Signal: A deep model begins producing NaN loss midway through training and the team increases epochs instead of checking gradient stability. Operational setting: At the model-governance gate before promotion of a new ML artifact.

D:

Compare RMSE with MAE or robust error summaries and choose the metric matching the application's penalty for large misses. This selection concerns rmse versus mae outlier sensitivity rather than the required diagnostic-and-remedy pair for model debugger exploding gradient diagnosis. Use evidence: Inspect tensor summaries, gradient magnitudes, learning-rate schedule, and training loss anomalies with supported debugging hooks. and the matching correction: Use appropriate SageMaker Debugger or framework instrumentation; adjust learning rate or gradient clipping after confirming instability. Signal: A deep model begins producing NaN loss midway through training and the team increases epochs instead of checking gradient stability. Operational setting: At the model-governance gate before promotion of a new ML artifact.

E:

Select a licensed, task-appropriate model and validate its performance and resource needs before a controlled deployment. This selection concerns sagemaker jumpstart template provenance rather than the required diagnostic-and-remedy pair for model debugger exploding gradient diagnosis. Use evidence: Inspect tensor summaries, gradient magnitudes, learning-rate schedule, and training loss anomalies with supported debugging hooks. and the matching correction: Use appropriate SageMaker Debugger or framework instrumentation; adjust learning rate or gradient clipping after confirming instability. Signal: A deep model begins producing NaN loss midway through training and the team increases epochs instead of checking gradient stability. Operational setting: At the model-governance gate before promotion of a new ML artifact.

Question 1184

While testing a second region for recoverable deployment of the same ML service, the following condition is observed: An edge vision model is deployed to a constrained device with a framework binary and operators assume any model artifact runs on its target accelerator. The deployment must be reproducible without manual console-only fixes. Which TWO steps best establish the evidence and correct the identified root cause? Select TWO.

  1. Compile the supported model for the specific edge target using SageMaker Neo or choose a compatible conversion route.
  2. Inspect target processor, supported Neo compiler configurations, operator compatibility, and deployment artifact format.
  3. Inspect deployment guardrail alarms, variant routing, old configuration retention, and rollback tests.
  4. Use a compatible SageMaker multi-model endpoint with managed model loading and monitor cache/eviction behavior.
  5. Reference an approved immutable image digest or controlled version tag and retain a reproducible build record.

Explanation

Correct Answers

A:

Compile the supported model for the specific edge target using SageMaker Neo or choose a compatible conversion route. This is the appropriate controlled remedy for sagemaker neo edge compiler compatibility. Edge optimization depends on target hardware and supported operations, not merely copying a cloud training artifact. Verify the cause with: Inspect target processor, supported Neo compiler configurations, operator compatibility, and deployment artifact format. Then confirm: The compiled artifact runs on the target device with validated outputs and acceptable memory and latency. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An edge vision model is deployed to a constrained device with a framework binary and operators assume any model artifact runs on its target accelerator. Operational setting: While testing a second region for recoverable deployment of the same ML service.

B:

Inspect target processor, supported Neo compiler configurations, operator compatibility, and deployment artifact format. This directly tests sagemaker neo edge compiler compatibility in the reported AWS ML workflow. Edge optimization depends on target hardware and supported operations, not merely copying a cloud training artifact. It distinguishes the problem before choosing a change. The related targeted response is: Compile the supported model for the specific edge target using SageMaker Neo or choose a compatible conversion route. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An edge vision model is deployed to a constrained device with a framework binary and operators assume any model artifact runs on its target accelerator. Operational setting: While testing a second region for recoverable deployment of the same ML service.

Incorrect Answers

C:

Inspect deployment guardrail alarms, variant routing, old configuration retention, and rollback tests. This selection concerns blue-green endpoint rollback after health failure rather than the required diagnostic-and-remedy pair for sagemaker neo edge compiler compatibility. Use evidence: Inspect target processor, supported Neo compiler configurations, operator compatibility, and deployment artifact format. and the matching correction: Compile the supported model for the specific edge target using SageMaker Neo or choose a compatible conversion route. Signal: An edge vision model is deployed to a constrained device with a framework binary and operators assume any model artifact runs on its target accelerator. Operational setting: While testing a second region for recoverable deployment of the same ML service.

D:

Use a compatible SageMaker multi-model endpoint with managed model loading and monitor cache/eviction behavior. This selection concerns multi-model endpoint hosting tradeoff rather than the required diagnostic-and-remedy pair for sagemaker neo edge compiler compatibility. Use evidence: Inspect target processor, supported Neo compiler configurations, operator compatibility, and deployment artifact format. and the matching correction: Compile the supported model for the specific edge target using SageMaker Neo or choose a compatible conversion route. Signal: An edge vision model is deployed to a constrained device with a framework binary and operators assume any model artifact runs on its target accelerator. Operational setting: While testing a second region for recoverable deployment of the same ML service.

E:

Reference an approved immutable image digest or controlled version tag and retain a reproducible build record. This selection concerns ecr immutable image digest for inference rather than the required diagnostic-and-remedy pair for sagemaker neo edge compiler compatibility. Use evidence: Inspect target processor, supported Neo compiler configurations, operator compatibility, and deployment artifact format. and the matching correction: Compile the supported model for the specific edge target using SageMaker Neo or choose a compatible conversion route. Signal: An edge vision model is deployed to a constrained device with a framework binary and operators assume any model artifact runs on its target accelerator. Operational setting: While testing a second region for recoverable deployment of the same ML service.

Question 1185

As a regulated ML workload prepares for an automated regional rollout, the following condition is observed: An A/B model test routes consecutive requests from the same user to different variants, contaminating the comparison with unstable user experience. The team must demonstrate a repeatable monitoring or security control. Which TWO complementary choices provide the decisive diagnosis AND corrective response? Select TWO.

  1. Inspect endpoint 5XXErrors, InvocationsPerInstance, instance availability, and per-minute latency distribution.
  2. Inspect routing hashes, experiment assignments, variant weights, and request identity consistency.
  3. Capture stable prediction identifiers and ingest verified labels with matching keys for model-quality analysis.
  4. Grant only the needed S3 and KMS actions for the approved data prefix and ensure appropriate role trust.
  5. Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable.

Explanation

Correct Answers

B:

Inspect routing hashes, experiment assignments, variant weights, and request identity consistency. This directly tests champion challenger a/b assignment consistency in the reported AWS ML workflow. Random assignment per request can mix treatments for one entity and bias longitudinal online experiments. It distinguishes the problem before choosing a change. The related targeted response is: Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An A/B model test routes consecutive requests from the same user to different variants, contaminating the comparison with unstable user experience. Operational setting: As a regulated ML workload prepares for an automated regional rollout.

E:

Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable. This is the appropriate controlled remedy for champion challenger a/b assignment consistency. Random assignment per request can mix treatments for one entity and bias longitudinal online experiments. Verify the cause with: Inspect routing hashes, experiment assignments, variant weights, and request identity consistency. Then confirm: Repeated requests for one test user reach the same assigned variant and measured outcomes remain comparable. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An A/B model test routes consecutive requests from the same user to different variants, contaminating the comparison with unstable user experience. Operational setting: As a regulated ML workload prepares for an automated regional rollout.

Incorrect Answers

A:

Inspect endpoint 5XXErrors, InvocationsPerInstance, instance availability, and per-minute latency distribution. This selection concerns cloudwatch 5xxerrors endpoint saturation rather than the required diagnostic-and-remedy pair for champion challenger a/b assignment consistency. Use evidence: Inspect routing hashes, experiment assignments, variant weights, and request identity consistency. and the matching correction: Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable. Signal: An A/B model test routes consecutive requests from the same user to different variants, contaminating the comparison with unstable user experience. Operational setting: As a regulated ML workload prepares for an automated regional rollout.

C:

Capture stable prediction identifiers and ingest verified labels with matching keys for model-quality analysis. This selection concerns model monitor ground truth association rather than the required diagnostic-and-remedy pair for champion challenger a/b assignment consistency. Use evidence: Inspect routing hashes, experiment assignments, variant weights, and request identity consistency. and the matching correction: Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable. Signal: An A/B model test routes consecutive requests from the same user to different variants, contaminating the comparison with unstable user experience. Operational setting: As a regulated ML workload prepares for an automated regional rollout.

D:

Grant only the needed S3 and KMS actions for the approved data prefix and ensure appropriate role trust. This selection concerns sagemaker execution role s3 least privilege rather than the required diagnostic-and-remedy pair for champion challenger a/b assignment consistency. Use evidence: Inspect routing hashes, experiment assignments, variant weights, and request identity consistency. and the matching correction: Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable. Signal: An A/B model test routes consecutive requests from the same user to different variants, contaminating the comparison with unstable user experience. Operational setting: As a regulated ML workload prepares for an automated regional rollout.

Question 1186

During production evaluation of a SageMaker AI model candidate, the following condition is observed: A credit-risk pilot selects a high-performing opaque model even though reviewers must justify material prediction features to an auditor. The release is blocked until the underlying model behavior is understood. Which observation most directly identifies the root cause of the reported behavior?

  1. Review model card, supported task, license terms, sample inference, evaluation evidence, and required compute footprint.
  2. Inspect documented explanation requirements, candidate feature contributions, and available post-hoc interpretation tests.
  3. Inspect experiment run lineage, dataset/version hashes, code revisions, random seeds, and metric definitions.
  4. Profile model memory, throughput, communication overhead, and per-device utilization.

Explanation

Correct Answer

B:

Inspect documented explanation requirements, candidate feature contributions, and available post-hoc interpretation tests. This directly tests interpretability for high-stakes prediction in the reported AWS ML workflow. Feature attribution can support explanation, but model choice must reflect transparency requirements and attribution is not causality. It distinguishes the problem before choosing a change. The related targeted response is: Compare a more interpretable model or add tested SageMaker Clarify explainability to the decision process. Reported signal: A credit-risk pilot selects a high-performing opaque model even though reviewers must justify material prediction features to an auditor. Operational setting: During production evaluation of a SageMaker AI model candidate.

Incorrect Answers

A:

Review model card, supported task, license terms, sample inference, evaluation evidence, and required compute footprint. This is a possible diagnostic for sagemaker jumpstart template provenance, but it does not directly test the interpretability for high-stakes prediction failure reported here. Prioritize: Inspect documented explanation requirements, candidate feature contributions, and available post-hoc interpretation tests. The relevant case is: A credit-risk pilot selects a high-performing opaque model even though reviewers must justify material prediction features to an auditor. Operational setting: During production evaluation of a SageMaker AI model candidate.

C:

Inspect experiment run lineage, dataset/version hashes, code revisions, random seeds, and metric definitions. This is a possible diagnostic for reproducible sagemaker experiments comparisons, but it does not directly test the interpretability for high-stakes prediction failure reported here. Prioritize: Inspect documented explanation requirements, candidate feature contributions, and available post-hoc interpretation tests. The relevant case is: A credit-risk pilot selects a high-performing opaque model even though reviewers must justify material prediction features to an auditor. Operational setting: During production evaluation of a SageMaker AI model candidate.

D:

Profile model memory, throughput, communication overhead, and per-device utilization. This is a possible diagnostic for data parallelism versus model parallelism, but it does not directly test the interpretability for high-stakes prediction failure reported here. Prioritize: Inspect documented explanation requirements, candidate feature contributions, and available post-hoc interpretation tests. The relevant case is: A credit-risk pilot selects a high-performing opaque model even though reviewers must justify material prediction features to an auditor. Operational setting: During production evaluation of a SageMaker AI model candidate.

Question 1187

During a controlled replay of the data-preparation workflow in a separate ML staging account, the following condition is observed: An overnight ML export runs expensive full-table queries on a production RDS database and causes application transaction latency spikes. The dataset contract must remain reproducible for a backfill. Which TWO complementary choices provide the decisive diagnosis AND corrective response? Select TWO.

  1. Choose partition keys with enough cardinality or reshard the stream for the intended throughput, while preserving required per-key ordering.
  2. Correlate export query plans, RDS CPU and IOPS, replication lag, and business traffic windows.
  3. Use controlled incremental extraction or a suitable read replica and throttle export jobs within the approved data freshness target.
  4. Inspect group sample sizes, label-positive fractions, and the appropriate SageMaker Clarify pretraining bias report.
  5. Encode hour with sine and cosine cyclic features and validate with a time-aware split.

Explanation

Correct Answers

B:

Correlate export query plans, RDS CPU and IOPS, replication lag, and business traffic windows. This directly tests rds extraction overload into ml in the reported AWS ML workflow. Extract jobs contend with transactional queries; moving or bounding reads can isolate production workloads. It distinguishes the problem before choosing a change. The related targeted response is: Use controlled incremental extraction or a suitable read replica and throttle export jobs within the approved data freshness target. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An overnight ML export runs expensive full-table queries on a production RDS database and causes application transaction latency spikes. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

C:

Use controlled incremental extraction or a suitable read replica and throttle export jobs within the approved data freshness target. This is the appropriate controlled remedy for rds extraction overload into ml. Extract jobs contend with transactional queries; moving or bounding reads can isolate production workloads. Verify the cause with: Correlate export query plans, RDS CPU and IOPS, replication lag, and business traffic windows. Then confirm: Application p95 latency returns to target while exported records meet a measured completeness and staleness objective. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An overnight ML export runs expensive full-table queries on a production RDS database and causes application transaction latency spikes. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

Incorrect Answers

A:

Choose partition keys with enough cardinality or reshard the stream for the intended throughput, while preserving required per-key ordering. This selection concerns kinesis data streams shard hot key rather than the required diagnostic-and-remedy pair for rds extraction overload into ml. Use evidence: Correlate export query plans, RDS CPU and IOPS, replication lag, and business traffic windows. and the matching correction: Use controlled incremental extraction or a suitable read replica and throttle export jobs within the approved data freshness target. Signal: An overnight ML export runs expensive full-table queries on a production RDS database and causes application transaction latency spikes. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

D:

Inspect group sample sizes, label-positive fractions, and the appropriate SageMaker Clarify pretraining bias report. This selection concerns clarify label imbalance measurement rather than the required diagnostic-and-remedy pair for rds extraction overload into ml. Use evidence: Correlate export query plans, RDS CPU and IOPS, replication lag, and business traffic windows. and the matching correction: Use controlled incremental extraction or a suitable read replica and throttle export jobs within the approved data freshness target. Signal: An overnight ML export runs expensive full-table queries on a production RDS database and causes application transaction latency spikes. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

E:

Encode hour with sine and cosine cyclic features and validate with a time-aware split. This selection concerns time-series cyclical calendar feature rather than the required diagnostic-and-remedy pair for rds extraction overload into ml. Use evidence: Correlate export query plans, RDS CPU and IOPS, replication lag, and business traffic windows. and the matching correction: Use controlled incremental extraction or a suitable read replica and throttle export jobs within the approved data freshness target. Signal: An overnight ML export runs expensive full-table queries on a production RDS database and causes application transaction latency spikes. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

Question 1188

During an operational load and release rehearsal for SageMaker-hosted inference, the following condition is observed: A pipeline pushes a newly trained model directly to production even though its evaluated recall has fallen below the business threshold. Change control requires measurable endpoint or pipeline acceptance evidence. Which observation most directly identifies the root cause of the reported behavior?

  1. Inspect the DAG of processing, training, evaluation and registration steps, along with parameter and artifact lineage.
  2. Inspect pipeline stages, evaluation artifacts, test result propagation, registry approval state, and deployment dependencies.
  3. Inspect model sizes, framework compatibility, per-model request rates, load latency, and isolation requirements.
  4. Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load.

Explanation

Correct Answer

B:

Inspect pipeline stages, evaluation artifacts, test result propagation, registry approval state, and deployment dependencies. This directly tests codepipeline approval gate on model metrics in the reported AWS ML workflow. A CI/CD stage that records evaluation without enforcing its result is not a release control. It distinguishes the problem before choosing a change. The related targeted response is: Require model evaluation and explicit approval gates before production deployment, blocking failed metric thresholds. Reported signal: A pipeline pushes a newly trained model directly to production even though its evaluated recall has fallen below the business threshold. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

Incorrect Answers

A:

Inspect the DAG of processing, training, evaluation and registration steps, along with parameter and artifact lineage. This is a possible diagnostic for sagemaker pipelines parameterized retraining, but it does not directly test the codepipeline approval gate on model metrics failure reported here. Prioritize: Inspect pipeline stages, evaluation artifacts, test result propagation, registry approval state, and deployment dependencies. The relevant case is: A pipeline pushes a newly trained model directly to production even though its evaluated recall has fallen below the business threshold. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

C:

Inspect model sizes, framework compatibility, per-model request rates, load latency, and isolation requirements. This is a possible diagnostic for multi-model endpoint hosting tradeoff, but it does not directly test the codepipeline approval gate on model metrics failure reported here. Prioritize: Inspect pipeline stages, evaluation artifacts, test result propagation, registry approval state, and deployment dependencies. The relevant case is: A pipeline pushes a newly trained model directly to production even though its evaluated recall has fallen below the business threshold. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

D:

Profile model operations, memory footprint, CPU saturation, request latency, and GPU throughput under representative load. This is a possible diagnostic for gpu versus cpu inference fit, but it does not directly test the codepipeline approval gate on model metrics failure reported here. Prioritize: Inspect pipeline stages, evaluation artifacts, test result propagation, registry approval state, and deployment dependencies. The relevant case is: A pipeline pushes a newly trained model directly to production even though its evaluated recall has fallen below the business threshold. Operational setting: During an operational load and release rehearsal for SageMaker-hosted inference.

Question 1189

During a production MLOps reliability and security review, the following condition is observed: A recommendation endpoint drops peak requests with HTTP 5xx as invocations per instance rise, while the dashboard averages latency across an entire day. The owner needs an auditable resolution without broadening privileges. Which engineering response resolves the actual issue without masking the symptoms?

  1. Set budget alerts and bounded tuning job quotas before launching experiments, and enforce an owner-approved stopping policy.
  2. Configure actionable alarms, capacity, and supported scaling or backpressure based on the burst window.
  3. Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable.
  4. Restrict writes to approved bucket prefixes and apply appropriate resource and organizational guardrails.

Explanation

Correct Answer

B:

Configure actionable alarms, capacity, and supported scaling or backpressure based on the burst window. This is the appropriate controlled remedy for cloudwatch 5xxerrors endpoint saturation. Daily averages can conceal short capacity failures; monitoring must use an incident-relevant interval. Verify the cause with: Inspect endpoint 5XXErrors, InvocationsPerInstance, instance availability, and per-minute latency distribution. Then confirm: Peak-window 5xx rate and p95 latency fall below service objectives with verified alarms. Reported signal: A recommendation endpoint drops peak requests with HTTP 5xx as invocations per instance rise, while the dashboard averages latency across an entire day. Operational setting: During a production MLOps reliability and security review.

Incorrect Answers

A:

Set budget alerts and bounded tuning job quotas before launching experiments, and enforce an owner-approved stopping policy. This addresses a different issue involving aws budgets spend threshold for experimentation. It does not correct the stated cloudwatch 5xxerrors endpoint saturation mechanism. The targeted response is: Configure actionable alarms, capacity, and supported scaling or backpressure based on the burst window. Case evidence: A recommendation endpoint drops peak requests with HTTP 5xx as invocations per instance rise, while the dashboard averages latency across an entire day. Operational setting: During a production MLOps reliability and security review.

C:

Use stable user-cohort assignment or an appropriate experiment design while keeping exposed groups measurable. This addresses a different issue involving champion challenger a/b assignment consistency. It does not correct the stated cloudwatch 5xxerrors endpoint saturation mechanism. The targeted response is: Configure actionable alarms, capacity, and supported scaling or backpressure based on the burst window. Case evidence: A recommendation endpoint drops peak requests with HTTP 5xx as invocations per instance rise, while the dashboard averages latency across an entire day. Operational setting: During a production MLOps reliability and security review.

D:

Restrict writes to approved bucket prefixes and apply appropriate resource and organizational guardrails. This addresses a different issue involving cross-account s3 bucket exfiltration protection. It does not correct the stated cloudwatch 5xxerrors endpoint saturation mechanism. The targeted response is: Configure actionable alarms, capacity, and supported scaling or backpressure based on the burst window. Case evidence: A recommendation endpoint drops peak requests with HTTP 5xx as invocations per instance rise, while the dashboard averages latency across an entire day. Operational setting: During a production MLOps reliability and security review.

Question 1190

During a controlled replay of the data-preparation workflow in a separate ML staging account, the following condition is observed: An offline churn dataset uses the latest cancellation-status feature for all historical rows, exposing outcomes not known when earlier predictions would run. The dataset contract must remain reproducible for a backfill. Which TWO complementary choices provide the decisive diagnosis AND corrective response? Select TWO.

  1. Enforce a required non-null or uniqueness contract with thresholds; quarantine violating partitions before feature publication.
  2. Define event-time windows with appropriate lateness handling and a documented revision policy for late records.
  3. Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value.
  4. Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query.
  5. Inspect sample raw records, quoting rules, expected headers, and parsed column counts in the ingestion job.

Explanation

Correct Answers

C:

Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value. This is the appropriate controlled remedy for point-in-time feature data leakage. A feature created after the prediction event cannot legitimately inform that earlier prediction. Verify the cause with: Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query. Then confirm: For every training row the joined feature timestamp is at or before its prediction timestamp and no label-derived value leaks. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An offline churn dataset uses the latest cancellation-status feature for all historical rows, exposing outcomes not known when earlier predictions would run. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

D:

Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query. This directly tests point-in-time feature data leakage in the reported AWS ML workflow. A feature created after the prediction event cannot legitimately inform that earlier prediction. It distinguishes the problem before choosing a change. The related targeted response is: Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: An offline churn dataset uses the latest cancellation-status feature for all historical rows, exposing outcomes not known when earlier predictions would run. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

Incorrect Answers

A:

Enforce a required non-null or uniqueness contract with thresholds; quarantine violating partitions before feature publication. This selection concerns glue data quality non-null contract rather than the required diagnostic-and-remedy pair for point-in-time feature data leakage. Use evidence: Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query. and the matching correction: Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value. Signal: An offline churn dataset uses the latest cancellation-status feature for all historical rows, exposing outcomes not known when earlier predictions would run. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

B:

Define event-time windows with appropriate lateness handling and a documented revision policy for late records. This selection concerns streaming feature rolling window correctness rather than the required diagnostic-and-remedy pair for point-in-time feature data leakage. Use evidence: Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query. and the matching correction: Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value. Signal: An offline churn dataset uses the latest cancellation-status feature for all historical rows, exposing outcomes not known when earlier predictions would run. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

E:

Inspect sample raw records, quoting rules, expected headers, and parsed column counts in the ingestion job. This selection concerns csv schema ambiguity during ingestion rather than the required diagnostic-and-remedy pair for point-in-time feature data leakage. Use evidence: Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query. and the matching correction: Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value. Signal: An offline churn dataset uses the latest cancellation-status feature for all historical rows, exposing outcomes not known when earlier predictions would run. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.

AWS Certified Machine Learning Engineer - Associate MLA-C01 Practice Tests