Amazon AWS Certified Machine Learning Engineer - Associate MLA-C01 Practice Test 86
Question 850
As a regional data-pipeline rollout is validated against approved AWS storage and streams, the following condition is observed: A rare-defect classifier oversamples minority instances before creating cross-validation folds, causing near copies in both training and holdout. The data team must preserve event-time meaning and lineage. What targeted change should be implemented after verifying the stated cause?
- Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value.
- Increase appropriate retention before long maintenance and persist a durable downstream historical copy for replay beyond retention.
- Fit an appropriate scaler on training data and package the same transformation with inference preprocessing.
- Apply resampling only inside each training fold and evaluate untouched validation folds using precision and recall metrics.
Explanation
Correct Answer
D:Apply resampling only inside each training fold and evaluate untouched validation folds using precision and recall metrics. This is the appropriate controlled remedy for class imbalance handled inside cross-validation. Oversampling before splitting contaminates evaluation and hides the real error rate on rare classes. Verify the cause with: Inspect sampler placement, duplicate neighbor IDs, fold membership, and class-conditioned metric changes. Then confirm: No synthetic or duplicated minority instance leaks into held-out folds, and minority-class recall reflects genuine unseen cases. Reported signal: A rare-defect classifier oversamples minority instances before creating cross-validation folds, causing near copies in both training and holdout. Operational setting: As a regional data-pipeline rollout is validated against approved AWS storage and streams.
Incorrect Answers
A:Perform point-in-time correct joins using the most recent eligible pre-prediction feature record, not the globally latest value. This addresses a different issue involving point-in-time feature data leakage. It does not correct the stated class imbalance handled inside cross-validation mechanism. The targeted response is: Apply resampling only inside each training fold and evaluate untouched validation folds using precision and recall metrics. Case evidence: A rare-defect classifier oversamples minority instances before creating cross-validation folds, causing near copies in both training and holdout. Operational setting: As a regional data-pipeline rollout is validated against approved AWS storage and streams.
B:Increase appropriate retention before long maintenance and persist a durable downstream historical copy for replay beyond retention. This addresses a different issue involving kinesis consumer retention gap. It does not correct the stated class imbalance handled inside cross-validation mechanism. The targeted response is: Apply resampling only inside each training fold and evaluate untouched validation folds using precision and recall metrics. Case evidence: A rare-defect classifier oversamples minority instances before creating cross-validation folds, causing near copies in both training and holdout. Operational setting: As a regional data-pipeline rollout is validated against approved AWS storage and streams.
C:Fit an appropriate scaler on training data and package the same transformation with inference preprocessing. This addresses a different issue involving standardization of numerical features. It does not correct the stated class imbalance handled inside cross-validation mechanism. The targeted response is: Apply resampling only inside each training fold and evaluate untouched validation folds using precision and recall metrics. Case evidence: A rare-defect classifier oversamples minority instances before creating cross-validation folds, causing near copies in both training and holdout. Operational setting: As a regional data-pipeline rollout is validated against approved AWS storage and streams.
Question 851
As a regulated ML workload prepares for an automated regional rollout, the following condition is observed: An endpoint's client-side response time rises, but the team only examines SageMaker ModelLatency and concludes the model is slow. The team must demonstrate a repeatable monitoring or security control. Which corrective action most directly addresses this ML workflow problem?
- Identify the component contributing to total request time and optimize network or overhead separately from actual model inference.
- Capture stable prediction identifiers and ingest verified labels with matching keys for model-quality analysis.
- Use SageMaker Inference Recommender or equivalent controlled load tests to rightsize the serving instance family.
- Use appropriate SageMaker Clarify model-bias monitoring and investigate supported group differences before deployment changes.
Explanation
Correct Answer
A:Identify the component contributing to total request time and optimize network or overhead separately from actual model inference. This is the appropriate controlled remedy for cloudwatch invocations and modellatency distinction. Model compute time and overall endpoint latency are distinct signals; blaming model code without tracing overhead is premature. Verify the cause with: Compare ModelLatency with OverheadLatency, client network timing, invocation counts, and endpoint 5XX responses. Then confirm: The chosen fix reduces measured end-to-end p95 latency and component metrics explain the change. Reported signal: An endpoint's client-side response time rises, but the team only examines SageMaker ModelLatency and concludes the model is slow. Operational setting: As a regulated ML workload prepares for an automated regional rollout.
Incorrect Answers
B:Capture stable prediction identifiers and ingest verified labels with matching keys for model-quality analysis. This addresses a different issue involving model monitor ground truth association. It does not correct the stated cloudwatch invocations and modellatency distinction mechanism. The targeted response is: Identify the component contributing to total request time and optimize network or overhead separately from actual model inference. Case evidence: An endpoint's client-side response time rises, but the team only examines SageMaker ModelLatency and concludes the model is slow. Operational setting: As a regulated ML workload prepares for an automated regional rollout.
C:Use SageMaker Inference Recommender or equivalent controlled load tests to rightsize the serving instance family. This addresses a different issue involving inference recommender benchmarking instance family. It does not correct the stated cloudwatch invocations and modellatency distinction mechanism. The targeted response is: Identify the component contributing to total request time and optimize network or overhead separately from actual model inference. Case evidence: An endpoint's client-side response time rises, but the team only examines SageMaker ModelLatency and concludes the model is slow. Operational setting: As a regulated ML workload prepares for an automated regional rollout.
D:Use appropriate SageMaker Clarify model-bias monitoring and investigate supported group differences before deployment changes. This addresses a different issue involving clarify postdeployment group bias monitoring. It does not correct the stated cloudwatch invocations and modellatency distinction mechanism. The targeted response is: Identify the component contributing to total request time and optimize network or overhead separately from actual model inference. Case evidence: An endpoint's client-side response time rises, but the team only examines SageMaker ModelLatency and concludes the model is slow. Operational setting: As a regulated ML workload prepares for an automated regional rollout.
Question 852
During production evaluation of a SageMaker AI model candidate, the following condition is observed: A tabular model shows one post-outcome administrative flag as its strongest feature in SageMaker Clarify despite implausibly high offline scores. The release is blocked until the underlying model behavior is understood. Which investigation provides the most decisive FIRST evidence for the underlying problem?
- Inspect confusion matrix, class prevalence, false-negative cost, and precision-recall curves across thresholds.
- Inspect model package version, artifact URI, training metadata, data snapshot, and registry approval state.
- Inspect Clarify attribution results, feature timestamp semantics, and whether the flag is created after the target event.
- Inspect training job InputDataConfig, environment variables such as SM_CHANNEL_TRAINING, and container paths.
Explanation
Correct Answer
C:Inspect Clarify attribution results, feature timestamp semantics, and whether the flag is created after the target event. This directly tests clarify feature attribution against target leakage in the reported AWS ML workflow. Attribution can expose suspicious shortcuts; a high importance value is not proof a feature is valid or causal. It distinguishes the problem before choosing a change. The related targeted response is: Remove label-derived or post-outcome features, rebuild a point-in-time split, and recompute explanations and holdout metrics. Reported signal: A tabular model shows one post-outcome administrative flag as its strongest feature in SageMaker Clarify despite implausibly high offline scores. Operational setting: During production evaluation of a SageMaker AI model candidate.
Incorrect Answers
A:Inspect confusion matrix, class prevalence, false-negative cost, and precision-recall curves across thresholds. This is a possible diagnostic for precision versus recall fraud threshold, but it does not directly test the clarify feature attribution against target leakage failure reported here. Prioritize: Inspect Clarify attribution results, feature timestamp semantics, and whether the flag is created after the target event. The relevant case is: A tabular model shows one post-outcome administrative flag as its strongest feature in SageMaker Clarify despite implausibly high offline scores. Operational setting: During production evaluation of a SageMaker AI model candidate.
B:Inspect model package version, artifact URI, training metadata, data snapshot, and registry approval state. This is a possible diagnostic for sagemaker model registry candidate approval, but it does not directly test the clarify feature attribution against target leakage failure reported here. Prioritize: Inspect Clarify attribution results, feature timestamp semantics, and whether the flag is created after the target event. The relevant case is: A tabular model shows one post-outcome administrative flag as its strongest feature in SageMaker Clarify despite implausibly high offline scores. Operational setting: During production evaluation of a SageMaker AI model candidate.
D:Inspect training job InputDataConfig, environment variables such as SM_CHANNEL_TRAINING, and container paths. This is a possible diagnostic for sagemaker script-mode data channel mismatch, but it does not directly test the clarify feature attribution against target leakage failure reported here. Prioritize: Inspect Clarify attribution results, feature timestamp semantics, and whether the flag is created after the target event. The relevant case is: A tabular model shows one post-outcome administrative flag as its strongest feature in SageMaker Clarify despite implausibly high offline scores. Operational setting: During production evaluation of a SageMaker AI model candidate.
Question 853
At a production AWS ML data lake supporting customer-facing predictions, the following condition is observed: An S3 event-triggered feature ingestion function processes the same uploaded object twice, creating duplicate labeled samples. Engineers must avoid changing unrelated source partitions. Which observation most directly identifies the root cause of the reported behavior?
- Inspect event identifiers, object versions or ETags, processing ledger, and downstream record keys for duplicate deliveries.
- Inspect per-class confusion, worker agreement, label instructions, and disagreements on an adjudicated reference set.
- Inspect feature scale distributions, model sensitivity, and whether fitted scaler statistics are shared across serving and training.
- Inspect workload read/write semantics, shared access requirements, filesystem operations, and object-store request patterns.
Explanation
Correct Answer
A:Inspect event identifiers, object versions or ETags, processing ledger, and downstream record keys for duplicate deliveries. This directly tests s3 object event at-least-once processing in the reported AWS ML workflow. Event notifications can be delivered more than once; consuming an event is not proof of exactly-once effects. It distinguishes the problem before choosing a change. The related targeted response is: Make the ingestion handler idempotent using a stable object/version key and conditional ledger writes. Reported signal: An S3 event-triggered feature ingestion function processes the same uploaded object twice, creating duplicate labeled samples. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
Incorrect Answers
B:Inspect per-class confusion, worker agreement, label instructions, and disagreements on an adjudicated reference set. This is a possible diagnostic for ground truth labeling agreement review, but it does not directly test the s3 object event at-least-once processing failure reported here. Prioritize: Inspect event identifiers, object versions or ETags, processing ledger, and downstream record keys for duplicate deliveries. The relevant case is: An S3 event-triggered feature ingestion function processes the same uploaded object twice, creating duplicate labeled samples. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
C:Inspect feature scale distributions, model sensitivity, and whether fitted scaler statistics are shared across serving and training. This is a possible diagnostic for standardization of numerical features, but it does not directly test the s3 object event at-least-once processing failure reported here. Prioritize: Inspect event identifiers, object versions or ETags, processing ledger, and downstream record keys for duplicate deliveries. The relevant case is: An S3 event-triggered feature ingestion function processes the same uploaded object twice, creating duplicate labeled samples. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
D:Inspect workload read/write semantics, shared access requirements, filesystem operations, and object-store request patterns. This is a possible diagnostic for s3 versus efs training storage choice, but it does not directly test the s3 object event at-least-once processing failure reported here. Prioritize: Inspect event identifiers, object versions or ETags, processing ledger, and downstream record keys for duplicate deliveries. The relevant case is: An S3 event-triggered feature ingestion function processes the same uploaded object twice, creating duplicate labeled samples. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
Question 854
At a production AWS ML data lake supporting customer-facing predictions, the following condition is observed: A demand forecast model is evaluated on randomly sampled days while features contain recent historical moving averages, creating unrealistic future information. Engineers must avoid changing unrelated source partitions. Which TWO steps best establish the evidence and correct the identified root cause? Select TWO.
- Use rolling-origin or chronological validation with strictly past-derived features and appropriate forecast horizons.
- Split by stable patient identity before training and reserve a disjoint held-out group for evaluation.
- Inspect feature scale distributions, model sensitivity, and whether fitted scaler statistics are shared across serving and training.
- Inspect feature availability timestamps and whether each label time follows the last feature observation.
- Write partitioned columnar Parquet data by event date, register accurate partitions in the catalog, and filter on partition keys.
Explanation
Correct Answers
A:Use rolling-origin or chronological validation with strictly past-derived features and appropriate forecast horizons. This is the appropriate controlled remedy for chronological split for forecast evaluation. Random splits violate temporal causality when future observations influence features or train data. Verify the cause with: Inspect feature availability timestamps and whether each label time follows the last feature observation. Then confirm: Every held-out forecast uses only data that existed at prediction time and its backtest follows real ordering. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: A demand forecast model is evaluated on randomly sampled days while features contain recent historical moving averages, creating unrealistic future information. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
D:Inspect feature availability timestamps and whether each label time follows the last feature observation. This directly tests chronological split for forecast evaluation in the reported AWS ML workflow. Random splits violate temporal causality when future observations influence features or train data. It distinguishes the problem before choosing a change. The related targeted response is: Use rolling-origin or chronological validation with strictly past-derived features and appropriate forecast horizons. This is one part of a select-two decision; the matching evidence and corrective action are both required. Reported signal: A demand forecast model is evaluated on randomly sampled days while features contain recent historical moving averages, creating unrealistic future information. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
Incorrect Answers
B:Split by stable patient identity before training and reserve a disjoint held-out group for evaluation. This selection concerns group-leakage train and validation split rather than the required diagnostic-and-remedy pair for chronological split for forecast evaluation. Use evidence: Inspect feature availability timestamps and whether each label time follows the last feature observation. and the matching correction: Use rolling-origin or chronological validation with strictly past-derived features and appropriate forecast horizons. Signal: A demand forecast model is evaluated on randomly sampled days while features contain recent historical moving averages, creating unrealistic future information. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
C:Inspect feature scale distributions, model sensitivity, and whether fitted scaler statistics are shared across serving and training. This selection concerns standardization of numerical features rather than the required diagnostic-and-remedy pair for chronological split for forecast evaluation. Use evidence: Inspect feature availability timestamps and whether each label time follows the last feature observation. and the matching correction: Use rolling-origin or chronological validation with strictly past-derived features and appropriate forecast horizons. Signal: A demand forecast model is evaluated on randomly sampled days while features contain recent historical moving averages, creating unrealistic future information. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
E:Write partitioned columnar Parquet data by event date, register accurate partitions in the catalog, and filter on partition keys. This selection concerns partition pruning in s3 parquet training input rather than the required diagnostic-and-remedy pair for chronological split for forecast evaluation. Use evidence: Inspect feature availability timestamps and whether each label time follows the last feature observation. and the matching correction: Use rolling-origin or chronological validation with strictly past-derived features and appropriate forecast horizons. Signal: A demand forecast model is evaluated on randomly sampled days while features contain recent historical moving averages, creating unrealistic future information. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
Question 855
While testing a second region for recoverable deployment of the same ML service, the following condition is observed: A real-time ML service is configured for very low baseline capacity but traffic can surge tenfold within seconds, exhausting available instance capacity. The deployment must be reproducible without manual console-only fixes. What postchange observation provides the strongest acceptance evidence?
- The overnight batch completes within its service window while there is no idle always-on inference endpoint charge.
- A synthetic load surge meets error and latency SLOs without rejected provisioning due to quota or startup lag.
- Low-frequency requests succeed within agreed latency while unused provisioned compute spend is reduced.
- Both regions match their approved parameterized templates and subsequent deployments remain repeatable.
Explanation
Correct Answer
B:A synthetic load surge meets error and latency SLOs without rejected provisioning due to quota or startup lag. This is the acceptance evidence for the described endpoint scale-to-zero cold capacity planning scenario. Reactive scaling takes time and is bounded by quotas; zero spare capacity may not absorb an instantaneous burst. The corrective implementation it validates is: Set adequate minimum capacity, scaling limits, alarms, and request backpressure for the expected traffic ramp. Reported signal: A real-time ML service is configured for very low baseline capacity but traffic can surge tenfold within seconds, exhausting available instance capacity. Operational setting: While testing a second region for recoverable deployment of the same ML service.
Incorrect Answers
A:The overnight batch completes within its service window while there is no idle always-on inference endpoint charge. This could validate real-time versus batch inference decision, but it does not establish that endpoint scale-to-zero cold capacity planning is resolved for this workload. Confirm instead: A synthetic load surge meets error and latency SLOs without rejected provisioning due to quota or startup lag. Observed problem: A real-time ML service is configured for very low baseline capacity but traffic can surge tenfold within seconds, exhausting available instance capacity. Operational setting: While testing a second region for recoverable deployment of the same ML service.
C:Low-frequency requests succeed within agreed latency while unused provisioned compute spend is reduced. This could validate serverless inference for intermittent requests, but it does not establish that endpoint scale-to-zero cold capacity planning is resolved for this workload. Confirm instead: A synthetic load surge meets error and latency SLOs without rejected provisioning due to quota or startup lag. Observed problem: A real-time ML service is configured for very low baseline capacity but traffic can surge tenfold within seconds, exhausting available instance capacity. Operational setting: While testing a second region for recoverable deployment of the same ML service.
D:Both regions match their approved parameterized templates and subsequent deployments remain repeatable. This could validate cloudformation stack parameter drift, but it does not establish that endpoint scale-to-zero cold capacity planning is resolved for this workload. Confirm instead: A synthetic load surge meets error and latency SLOs without rejected provisioning due to quota or startup lag. Observed problem: A real-time ML service is configured for very low baseline capacity but traffic can surge tenfold within seconds, exhausting available instance capacity. Operational setting: While testing a second region for recoverable deployment of the same ML service.
Question 856
During a controlled replay of the data-preparation workflow in a separate ML staging account, the following condition is observed: A model trained on transaction amount and account age is unstable because one feature spans millions of units and the other spans years. The dataset contract must remain reproducible for a backfill. What postchange observation provides the strongest acceptance evidence?
- Long-tail task duration and shuffle spill decline without changing the joined record counts.
- Training converges more consistently, and the same raw request gives aligned standardized values online.
- Write throttles and iterator age fall under burst load without losing required ordering for related records.
- Application p95 latency returns to target while exported records meet a measured completeness and staleness objective.
Explanation
Correct Answer
B:Training converges more consistently, and the same raw request gives aligned standardized values online. This is the acceptance evidence for the described standardization of numerical features scenario. Gradient-based training is sensitive to disparate feature scales; training-serving transformation parity is required. The corrective implementation it validates is: Fit an appropriate scaler on training data and package the same transformation with inference preprocessing. Reported signal: A model trained on transaction amount and account age is unstable because one feature spans millions of units and the other spans years. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
Incorrect Answers
A:Long-tail task duration and shuffle spill decline without changing the joined record counts. This could validate glue spark join skew on dominant identifier, but it does not establish that standardization of numerical features is resolved for this workload. Confirm instead: Training converges more consistently, and the same raw request gives aligned standardized values online. Observed problem: A model trained on transaction amount and account age is unstable because one feature spans millions of units and the other spans years. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
C:Write throttles and iterator age fall under burst load without losing required ordering for related records. This could validate kinesis data streams shard hot key, but it does not establish that standardization of numerical features is resolved for this workload. Confirm instead: Training converges more consistently, and the same raw request gives aligned standardized values online. Observed problem: A model trained on transaction amount and account age is unstable because one feature spans millions of units and the other spans years. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
D:Application p95 latency returns to target while exported records meet a measured completeness and staleness objective. This could validate rds extraction overload into ml, but it does not establish that standardization of numerical features is resolved for this workload. Confirm instead: Training converges more consistently, and the same raw request gives aligned standardized values online. Observed problem: A model trained on transaction amount and account age is unstable because one feature spans millions of units and the other spans years. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
Question 857
During a controlled replay of the data-preparation workflow in a separate ML staging account, the following condition is observed: A near-real-time account feature sums transactions using processing time, so late-arriving events land in the wrong window and produce inconsistent scores. The dataset contract must remain reproducible for a backfill. Which investigation provides the most decisive FIRST evidence for the underlying problem?
- Inspect event timestamps, ingest timestamps, watermark configuration, and window outputs for delayed sample records.
- Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query.
- Compare vocabulary hashes, normalization steps, token IDs for test strings, and serialized preprocessing versions.
- Inspect shared entity identifiers across partitions and compare patient-level versus image-level evaluation results.
Explanation
Correct Answer
A:Inspect event timestamps, ingest timestamps, watermark configuration, and window outputs for delayed sample records. This directly tests streaming feature rolling window correctness in the reported AWS ML workflow. Processing-time windows depend on arrival timing; event-time windows need explicit watermarks and late-event policy. It distinguishes the problem before choosing a change. The related targeted response is: Define event-time windows with appropriate lateness handling and a documented revision policy for late records. Reported signal: A near-real-time account feature sums transactions using processing time, so late-arriving events land in the wrong window and produce inconsistent scores. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
Incorrect Answers
B:Inspect feature event timestamps, target prediction cutoffs, and join predicates in the historical training query. This is a possible diagnostic for point-in-time feature data leakage, but it does not directly test the streaming feature rolling window correctness failure reported here. Prioritize: Inspect event timestamps, ingest timestamps, watermark configuration, and window outputs for delayed sample records. The relevant case is: A near-real-time account feature sums transactions using processing time, so late-arriving events land in the wrong window and produce inconsistent scores. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
C:Compare vocabulary hashes, normalization steps, token IDs for test strings, and serialized preprocessing versions. This is a possible diagnostic for text tokenization mismatch in serving, but it does not directly test the streaming feature rolling window correctness failure reported here. Prioritize: Inspect event timestamps, ingest timestamps, watermark configuration, and window outputs for delayed sample records. The relevant case is: A near-real-time account feature sums transactions using processing time, so late-arriving events land in the wrong window and produce inconsistent scores. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
D:Inspect shared entity identifiers across partitions and compare patient-level versus image-level evaluation results. This is a possible diagnostic for group-leakage train and validation split, but it does not directly test the streaming feature rolling window correctness failure reported here. Prioritize: Inspect event timestamps, ingest timestamps, watermark configuration, and window outputs for delayed sample records. The relevant case is: A near-real-time account feature sums transactions using processing time, so late-arriving events land in the wrong window and produce inconsistent scores. Operational setting: During a controlled replay of the data-preparation workflow in a separate ML staging account.
Question 858
During a production MLOps reliability and security review, the following condition is observed: An inference container embeds a database password in its image, and every rollout reuses the credential even after an incident. The owner needs an auditable resolution without broadening privileges. What targeted change should be implemented after verifying the stated cause?
- Identify the component contributing to total request time and optimize network or overhead separately from actual model inference.
- Remove embedded credentials, rotate the exposed secret, and retrieve current credentials through an approved managed secret service.
- Apply consistent approved cost allocation tags to ML jobs and supported linked resources and activate relevant reporting.
- Use approved SageMaker interface endpoints and relevant private DNS and policy settings for required API access.
Explanation
Correct Answer
B:Remove embedded credentials, rotate the exposed secret, and retrieve current credentials through an approved managed secret service. This is the appropriate controlled remedy for secret rotation for inference container credentials. Container artifacts can leak hard-coded secrets; runtime retrieval and rotation keep sensitive values outside build outputs. Verify the cause with: Inspect image layers, environment configuration, secret source, runtime role permissions, and rotation logs. Then confirm: The new container operates with rotated credentials and image scans find no active embedded password. Reported signal: An inference container embeds a database password in its image, and every rollout reuses the credential even after an incident. Operational setting: During a production MLOps reliability and security review.
Incorrect Answers
A:Identify the component contributing to total request time and optimize network or overhead separately from actual model inference. This addresses a different issue involving cloudwatch invocations and modellatency distinction. It does not correct the stated secret rotation for inference container credentials mechanism. The targeted response is: Remove embedded credentials, rotate the exposed secret, and retrieve current credentials through an approved managed secret service. Case evidence: An inference container embeds a database password in its image, and every rollout reuses the credential even after an incident. Operational setting: During a production MLOps reliability and security review.
C:Apply consistent approved cost allocation tags to ML jobs and supported linked resources and activate relevant reporting. This addresses a different issue involving tag-based ml training cost attribution. It does not correct the stated secret rotation for inference container credentials mechanism. The targeted response is: Remove embedded credentials, rotate the exposed secret, and retrieve current credentials through an approved managed secret service. Case evidence: An inference container embeds a database password in its image, and every rollout reuses the credential even after an incident. Operational setting: During a production MLOps reliability and security review.
D:Use approved SageMaker interface endpoints and relevant private DNS and policy settings for required API access. This addresses a different issue involving private vpc endpoint policy for sagemaker apis. It does not correct the stated secret rotation for inference container credentials mechanism. The targeted response is: Remove embedded credentials, rotate the exposed secret, and retrieve current credentials through an approved managed secret service. Case evidence: An inference container embeds a database password in its image, and every rollout reuses the credential even after an incident. Operational setting: During a production MLOps reliability and security review.
Question 859
At a production AWS ML data lake supporting customer-facing predictions, the following condition is observed: A hiring-model dataset has very different positive-label rates across sensitive groups, but the team compares only global accuracy. Engineers must avoid changing unrelated source partitions. Which verification result BEST demonstrates that the issue is resolved?
- The report quantifies class or group imbalance with reproducible group counts and documents remediation assumptions.
- Validated records preserve the expected label and field order while malformed inputs are reported separately.
- A test batch remains in the approved region with enforceable policy and audit evidence for all derived artifacts.
- The transformed distribution is less skewed and inference uses identical guarded transformations.
Explanation
Correct Answer
A:The report quantifies class or group imbalance with reproducible group counts and documents remediation assumptions. This is the acceptance evidence for the described clarify label imbalance measurement scenario. Aggregate accuracy can conceal unequal source-label distributions, which pretraining metrics assess before model fitting. The corrective implementation it validates is: Measure label distribution differences with an appropriate pretraining bias metric and improve data representation before training. Reported signal: A hiring-model dataset has very different positive-label rates across sensitive groups, but the team compares only global accuracy. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
Incorrect Answers
B:Validated records preserve the expected label and field order while malformed inputs are reported separately. This could validate csv schema ambiguity during ingestion, but it does not establish that clarify label imbalance measurement is resolved for this workload. Confirm instead: The report quantifies class or group imbalance with reproducible group counts and documents remediation assumptions. Observed problem: A hiring-model dataset has very different positive-label rates across sensitive groups, but the team compares only global accuracy. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
C:A test batch remains in the approved region with enforceable policy and audit evidence for all derived artifacts. This could validate model training data residency boundary, but it does not establish that clarify label imbalance measurement is resolved for this workload. Confirm instead: The report quantifies class or group imbalance with reproducible group counts and documents remediation assumptions. Observed problem: A hiring-model dataset has very different positive-label rates across sensitive groups, but the team compares only global accuracy. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
D:The transformed distribution is less skewed and inference uses identical guarded transformations. This could validate skewed numerical feature log transformation, but it does not establish that clarify label imbalance measurement is resolved for this workload. Confirm instead: The report quantifies class or group imbalance with reproducible group counts and documents remediation assumptions. Observed problem: A hiring-model dataset has very different positive-label rates across sensitive groups, but the team compares only global accuracy. Operational setting: At a production AWS ML data lake supporting customer-facing predictions.
Question 860
While reproducing the alert in an isolated AWS ML integration account, the following condition is observed: An asynchronous inference queue's oldest pending request age keeps increasing, yet GPU utilization alerts remain normal because requests are waiting before execution. The proposed change must preserve logging and least privilege. What postchange observation provides the strongest acceptance evidence?
- Auditors see reproducible subgroup statistics and alarms for material disparities rather than only global totals.
- An authorized training read emits attributable S3 object-access events while alerting detects unexpected access.
- Measured training cost drops while interrupted test jobs resume from checkpoints within the delivery window.
- Backlog age stabilizes and requests complete within the agreed service-level window.
Explanation
Correct Answer
D:Backlog age stabilizes and requests complete within the agreed service-level window. This is the acceptance evidence for the described asynchronous inference queue backlog alert scenario. Queued work can miss response deadlines without saturating compute; queue age is its own performance signal. The corrective implementation it validates is: Scale compatible asynchronous inference capacity or tune workload and queue alarms for pending-request age. Reported signal: An asynchronous inference queue's oldest pending request age keeps increasing, yet GPU utilization alerts remain normal because requests are waiting before execution. Operational setting: While reproducing the alert in an isolated AWS ML integration account.
Incorrect Answers
A:Auditors see reproducible subgroup statistics and alarms for material disparities rather than only global totals. This could validate clarify postdeployment group bias monitoring, but it does not establish that asynchronous inference queue backlog alert is resolved for this workload. Confirm instead: Backlog age stabilizes and requests complete within the agreed service-level window. Observed problem: An asynchronous inference queue's oldest pending request age keeps increasing, yet GPU utilization alerts remain normal because requests are waiting before execution. Operational setting: While reproducing the alert in an isolated AWS ML integration account.
B:An authorized training read emits attributable S3 object-access events while alerting detects unexpected access. This could validate cloudtrail training data access audit coverage, but it does not establish that asynchronous inference queue backlog alert is resolved for this workload. Confirm instead: Backlog age stabilizes and requests complete within the agreed service-level window. Observed problem: An asynchronous inference queue's oldest pending request age keeps increasing, yet GPU utilization alerts remain normal because requests are waiting before execution. Operational setting: While reproducing the alert in an isolated AWS ML integration account.
C:Measured training cost drops while interrupted test jobs resume from checkpoints within the delivery window. This could validate spot training cost versus interruption recovery, but it does not establish that asynchronous inference queue backlog alert is resolved for this workload. Confirm instead: Backlog age stabilizes and requests complete within the agreed service-level window. Observed problem: An asynchronous inference queue's oldest pending request age keeps increasing, yet GPU utilization alerts remain normal because requests are waiting before execution. Operational setting: While reproducing the alert in an isolated AWS ML integration account.
