cert
cert-1
cert-2

Pass Google Professional Machine Learning Engineer 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
Professional Machine Learning Engineer Exam - Verified By Experts
Professional Machine Learning Engineer Premium Bundle
$29.98

Professional Machine Learning Engineer Premium Bundle

$79.99
$109.97
  • Premium File 402 Questions & Answers. Last update: Nov 13, 2026
  • Training Course 69 Video Lectures
  • Study Guide 376 Pages
 
$109.97
$79.99
accept 47 downloads in last 7 days
block-screenshots
Purchase Google Professional Machine Learning Engineer Exam Training Products Individually
Google Professional Machine Learning Engineer Premium File
Premium File 402 Q&A
$76.99$69.99
Google Professional Machine Learning Engineer Video Training Course
Training Course 69 Lectures
$21.99 $19.99
Google Professional Machine Learning Engineer PDF Study Guide
Study Guide 376 Pages
$21.99 $19.99
Google Professional Machine Learning Engineer Practice Test 114

Question 1130

During model-training validation on a Google Cloud experimental project, the following condition occurs: A weeklong custom training job uses preemptible capacity but saves checkpoints only to a temporary local container filesystem. The change should preserve a meaningful independent model-quality comparison. Which underlying engineering principle explains why the original approach fails?

  1. Preemptible execution can end abruptly; local-only state is not a durable recovery mechanism.
  2. Reduced numeric precision narrows representable ranges and can underflow or overflow without proper scaling.
  3. An accelerator can sit idle waiting for input even when its own memory and compute capacity are sufficient.
  4. Notebook filesystem locations are not portable contracts for managed container training jobs.

Explanation

Correct Answer

A:

Preemptible execution can end abruptly; local-only state is not a durable recovery mechanism. This explains the mechanism behind training checkpoint persistence across interruptions and why the described failure is possible. A discriminating investigation is: Inspect shutdown events, checkpoint directory, durable artifact transfers and restoration logic after simulated interruptions. The corresponding repair is: Persist versioned checkpoints in durable Cloud Storage and make the training code resume optimizer and model state.

Incorrect Answers

B:

Reduced numeric precision narrows representable ranges and can underflow or overflow without proper scaling. This principle describes mixed precision overflow diagnosis, which is not the cause of the observed training checkpoint persistence across interruptions failure. The controlling explanation is: Preemptible execution can end abruptly; local-only state is not a durable recovery mechanism.

C:

An accelerator can sit idle waiting for input even when its own memory and compute capacity are sufficient. This principle describes accelerator starvation from slow input pipeline, which is not the cause of the observed training checkpoint persistence across interruptions failure. The controlling explanation is: Preemptible execution can end abruptly; local-only state is not a durable recovery mechanism.

D:

Notebook filesystem locations are not portable contracts for managed container training jobs. This principle describes custom training input schema mismatch, which is not the cause of the observed training checkpoint persistence across interruptions failure. The controlling explanation is: Preemptible execution can end abruptly; local-only state is not a durable recovery mechanism.

Question 1131

During a production ML inference release on Google Cloud, the following condition occurs: A team deploys a model using a mutable storage path named production-latest and cannot reconstruct which evaluation approved its weights. The service must meet its agreed quality and latency requirements. Which targeted engineering change most directly addresses the failure?

  1. Implement the required serving contract and make the container report readiness only after loading model assets.
  2. Use asynchronous or batch processing with durable job status for oversized workloads, reserving online calls for bounded requests.
  3. Promote an explicitly approved model version from Agent Platform Model Registry and record deployment lineage.
  4. Preserve stable canonical identifier types end-to-end and backfill only the affected feature records.

Explanation

Correct Answer

C:

Promote an explicitly approved model version from Agent Platform Model Registry and record deployment lineage. This is the targeted corrective response for model registry version provenance. Mutable nicknames do not establish immutable model provenance for controlled production releases. Verify the diagnosis using: Review immutable artifact IDs, Model Registry versions, evaluation reports, approvals and deployed revision metadata. Confirm success using: The running endpoint resolves to a traceable evaluated package and cannot silently switch weights when an alias moves. This matters here because a team deploys a model using a mutable storage path named production-latest and cannot reconstruct which evaluation….

Incorrect Answers

A:

Implement the required serving contract and make the container report readiness only after loading model assets. This action addresses inference container health interface mismatch, not the failure in the described model registry version provenance case. The targeted correction is: Promote an explicitly approved model version from Agent Platform Model Registry and record deployment lineage. This matters here because a team deploys a model using a mutable storage path named production-latest and cannot reconstruct which evaluation….

B:

Use asynchronous or batch processing with durable job status for oversized workloads, reserving online calls for bounded requests. This action addresses online prediction timeout and batch fallback, not the failure in the described model registry version provenance case. The targeted correction is: Promote an explicitly approved model version from Agent Platform Model Registry and record deployment lineage.

D:

Preserve stable canonical identifier types end-to-end and backfill only the affected feature records. This action addresses feature online entity key mismatch, not the failure in the described model registry version provenance case. The targeted correction is: Promote an explicitly approved model version from Agent Platform Model Registry and record deployment lineage.

Question 1132

As the team tests a new governed AI release against adversarial and normal data, the following condition occurs: An assistant summarizes third-party documents and malicious attachments can contain hidden exfiltration instructions, with no content screening before model invocation. The fix must retain legitimate model usefulness and appropriate controls. Select TWO: the most relevant root-cause check and the specific observation that proves the fix worked.

  1. Inspect attachment ingestion, trust boundaries, malicious prompt examples and deployed Model Armor filter results.
  2. A controlled prompt-injection fixture is blocked or neutralized without disrupting ordinary allowed documents.
  3. The candidate meets subgroup harm thresholds on representative held-out data without hiding global tradeoffs.
  4. Examine explanation distributions, feature generation changes, protected attribute proxies and prediction outcomes by cohort.
  5. A failing canary halts and restores the healthy model before broader traffic exposure.

Explanation

Correct Answers

A:

Inspect attachment ingestion, trust boundaries, malicious prompt examples and deployed Model Armor filter results. This inspects evidence relevant to model armor untrusted content screening. Screening helps detect malicious instructions or sensitive content crossing the model trust boundary. After identifying the cause, the matching correction is: Add appropriate Model Armor screening to untrusted inputs and outputs and reject or quarantine policy-violating content.

B:

A controlled prompt-injection fixture is blocked or neutralized without disrupting ordinary allowed documents. This is specific acceptance evidence for model armor untrusted content screening. Screening helps detect malicious instructions or sensitive content crossing the model trust boundary. It confirms the controlled correction: Add appropriate Model Armor screening to untrusted inputs and outputs and reject or quarantine policy-violating content.

Incorrect Answers

C:

The candidate meets subgroup harm thresholds on representative held-out data without hiding global tradeoffs. This relates to model feature bias by subgroup but does not establish the requested before-and-after evidence. The actual diagnostic is: Inspect attachment ingestion, trust boundaries, malicious prompt examples and deployed Model Armor filter results. The specific acceptance result is: A controlled prompt-injection fixture is blocked or neutralized without disrupting ordinary allowed documents.

D:

Examine explanation distributions, feature generation changes, protected attribute proxies and prediction outcomes by cohort. This relates to feature attribution drift after deployment but does not establish the requested before-and-after evidence. The actual diagnostic is: Inspect attachment ingestion, trust boundaries, malicious prompt examples and deployed Model Armor filter results. The specific acceptance result is: A controlled prompt-injection fixture is blocked or neutralized without disrupting ordinary allowed documents.

E:

A failing canary halts and restores the healthy model before broader traffic exposure. This relates to production incident rollback based on metrics but does not establish the requested before-and-after evidence. The actual diagnostic is: Inspect attachment ingestion, trust boundaries, malicious prompt examples and deployed Model Armor filter results. The specific acceptance result is: A controlled prompt-injection fixture is blocked or neutralized without disrupting ordinary allowed documents.

Question 1133

During a staged restoration test at a secondary serving deployment, the following condition occurs: An industrial camera must classify frames while offline for hours, but its prototype requires every frame to reach a remote cloud endpoint. The team must prove an actual user-visible improvement without weakening access controls. Select TWO: the decisive diagnostic step and the targeted repair for this specific issue.

  1. Examine endpoint exposure settings, caller IAM, VPC routes, private connectivity and data egress requirements.
  2. Use asynchronous or batch processing with durable job status for oversized workloads, reserving online calls for bounded requests.
  3. Deploy a verified edge-compatible model for required offline decisions and sync summarized results when connectivity returns.
  4. Measure bandwidth, network availability, latency deadline, edge hardware limits and model compression compatibility.
  5. Test a supported post-training quantization or optimization path and release only if target quality remains acceptable.

Explanation

Correct Answers

C:

Deploy a verified edge-compatible model for required offline decisions and sync summarized results when connectivity returns. This is the targeted corrective response for edge versus cloud inference network constraints. A cloud-only inference architecture cannot satisfy hard offline and low-latency edge requirements. Verify the diagnosis using: Measure bandwidth, network availability, latency deadline, edge hardware limits and model compression compatibility. Confirm success using: The camera processes representative frames inside its latency budget during an intentional network outage. This matters here because during a staged restoration test at a secondary serving deployment, the following condition occurs: An industrial….

D:

Measure bandwidth, network availability, latency deadline, edge hardware limits and model compression compatibility. This inspects evidence relevant to edge versus cloud inference network constraints. A cloud-only inference architecture cannot satisfy hard offline and low-latency edge requirements. After identifying the cause, the matching correction is: Deploy a verified edge-compatible model for required offline decisions and sync summarized results when connectivity returns. This matters here because during a staged restoration test at a secondary serving deployment, the following condition occurs: An industrial….

Incorrect Answers

A:

Examine endpoint exposure settings, caller IAM, VPC routes, private connectivity and data egress requirements. This is an action or diagnostic relevant to private endpoint for regulated inference traffic, not one of the required two steps. In this scenario verify: Measure bandwidth, network availability, latency deadline, edge hardware limits and model compression compatibility. Then correct: Deploy a verified edge-compatible model for required offline decisions and sync summarized results when connectivity returns. This matters here because during a staged restoration test at a secondary serving deployment, the following condition occurs: An industrial….

B:

Use asynchronous or batch processing with durable job status for oversized workloads, reserving online calls for bounded requests. This is an action or diagnostic relevant to online prediction timeout and batch fallback, not one of the required two steps. In this scenario verify: Measure bandwidth, network availability, latency deadline, edge hardware limits and model compression compatibility. Then correct: Deploy a verified edge-compatible model for required offline decisions and sync summarized results when connectivity returns.

E:

Test a supported post-training quantization or optimization path and release only if target quality remains acceptable. This is an action or diagnostic relevant to inference optimization using quantization, not one of the required two steps. In this scenario verify: Measure bandwidth, network availability, latency deadline, edge hardware limits and model compression compatibility. Then correct: Deploy a verified edge-compatible model for required offline decisions and sync summarized results when connectivity returns.

Question 1134

During model-training validation on a Google Cloud experimental project, the following condition occurs: An imbalanced defect dataset is oversampled before train/test splitting, so near-duplicate synthetic records appear in both partitions. The change should preserve a meaningful independent model-quality comparison. Which evidence most directly establishes what is wrong with this AI or ML workflow?

  1. Inspect output schema, decoding configuration, structured-output support and malformed samples under representative prompts.
  2. Compare pretraining benchmark scores, domain validation curves, trainable parameters and fine-tuning learning-rate choices.
  3. Inspect sampling order, example hashes and class distributions in training and held-out folds.
  4. Check dataset licenses, provenance, consent policies, safety requirements and the model's supported tuning route.

Explanation

Correct Answer

C:

Inspect sampling order, example hashes and class distributions in training and held-out folds. This inspects evidence relevant to training class imbalance fold leakage. Resampling before splitting leaks information and makes minority-class performance appear artificially high. After identifying the cause, the matching correction is: Split raw records first and resample only inside training folds, leaving validation cohorts untouched. This matters here because during model-training validation on a Google Cloud experimental project, the following condition occurs: An….

Incorrect Answers

A:

Inspect output schema, decoding configuration, structured-output support and malformed samples under representative prompts. This examines constrained generative json outputs rather than the specific mechanism reported here. It is not the decisive first check for training class imbalance fold leakage. The relevant evidence is: Inspect sampling order, example hashes and class distributions in training and held-out folds.

B:

Compare pretraining benchmark scores, domain validation curves, trainable parameters and fine-tuning learning-rate choices. This examines fine-tuning catastrophic forgetting rather than the specific mechanism reported here. It is not the decisive first check for training class imbalance fold leakage. The relevant evidence is: Inspect sampling order, example hashes and class distributions in training and held-out folds.

D:

Check dataset licenses, provenance, consent policies, safety requirements and the model's supported tuning route. This examines fine-tuning image generator data rights rather than the specific mechanism reported here. It is not the decisive first check for training class imbalance fold leakage. The relevant evidence is: Inspect sampling order, example hashes and class distributions in training and held-out folds.

Question 1135

During a regulated MLOps pipeline deployment, the following condition occurs: A pipeline retrains after every minor quality fluctuation and immediately promotes each candidate, causing frequent unstable production swaps. The workflow must be repeatable and protect the last approved production model. Which result is the strongest postchange evidence of success?

  1. Staging and production runs report the same component digests and only documented configuration values differ.
  2. No-op periods create no unnecessary models and validated drift events produce candidates requiring approval.
  3. Small harmless fluctuations do not force releases and genuine regressions invoke a tested rollback or controlled retraining.
  4. One completed batch starts one intended pipeline run and unrelated object uploads do not trigger training.

Explanation

Correct Answer

C:

Small harmless fluctuations do not force releases and genuine regressions invoke a tested rollback or controlled retraining. This is specific acceptance evidence for canary failure feedback into ct policy. Continuous training needs statistically and operationally justified policy thresholds rather than reactive promotion of noise. It confirms the controlled correction: Use a retraining policy with significant trigger thresholds, independent evaluation and staged deployment approval.

Incorrect Answers

A:

Staging and production runs report the same component digests and only documented configuration values differ. This result could demonstrate success for pipeline environment parameter promotion but not for the stated canary failure feedback into ct policy issue. The direct test is: Small harmless fluctuations do not force releases and genuine regressions invoke a tested rollback or controlled retraining.

B:

No-op periods create no unnecessary models and validated drift events produce candidates requiring approval. This result could demonstrate success for data drift triggered retraining policy but not for the stated canary failure feedback into ct policy issue. The direct test is: Small harmless fluctuations do not force releases and genuine regressions invoke a tested rollback or controlled retraining.

D:

One completed batch starts one intended pipeline run and unrelated object uploads do not trigger training. This result could demonstrate success for eventbridge equivalent event trigger on storage but not for the stated canary failure feedback into ct policy issue. The direct test is: Small harmless fluctuations do not force releases and genuine regressions invoke a tested rollback or controlled retraining.

Question 1136

During a postdeployment AI safety and reliability review, the following condition occurs: A service account running an AI workload can read all analytics datasets and copy confidential training data to an unrelated storage bucket. The model owner must provide repeatable evidence rather than a cosmetic dashboard change. Which targeted engineering change most directly addresses the failure?

  1. Minimize or tokenize personal fields before generation and enforce access and retention controls on any necessary context.
  2. Adopt justified seasonal baselines or cohorts and tune alert thresholds without suppressing severe shifts.
  3. Present feature attributions as model-decision associations and use proper causal analysis before making intervention claims.
  4. Restrict dataset and bucket access by principal and resource; enforce approved egress controls and audit sensitive reads.

Explanation

Correct Answer

D:

Restrict dataset and bucket access by principal and resource; enforce approved egress controls and audit sensitive reads. This is the targeted corrective response for training data exfiltration through broad access. Least-privilege access and destination controls reduce unauthorized model-data disclosure risk. Verify the diagnosis using: Review IAM resource bindings, bucket access policies, egress paths and audit records for unauthorized data transfers. Confirm success using: The authorized pipeline still reads its intended data while attempts to export to unapproved destinations are denied. This matters here because a service account running an AI workload can read all analytics datasets and copy confidential training data to an….

Incorrect Answers

A:

Minimize or tokenize personal fields before generation and enforce access and retention controls on any necessary context. This action addresses sensitive identifiers leaked to llm prompts, not the failure in the described training data exfiltration through broad access case. The targeted correction is: Restrict dataset and bucket access by principal and resource; enforce approved egress controls and audit sensitive reads.

B:

Adopt justified seasonal baselines or cohorts and tune alert thresholds without suppressing severe shifts. This action addresses model monitoring false-alert seasonality, not the failure in the described training data exfiltration through broad access case. The targeted correction is: Restrict dataset and bucket access by principal and resource; enforce approved egress controls and audit sensitive reads.

C:

Present feature attributions as model-decision associations and use proper causal analysis before making intervention claims. This action addresses explainability versus causal claim, not the failure in the described training data exfiltration through broad access case. The targeted correction is: Restrict dataset and bucket access by principal and resource; enforce approved egress controls and audit sensitive reads. This matters here because a service account running an AI workload can read all analytics datasets and copy confidential training data to an….

Question 1137

During a staged restoration test at a secondary serving deployment, the following condition occurs: A public prediction API queues unbounded client retries during a traffic spike, creating a cascading latency and memory outage. The team must prove an actual user-visible improvement without weakening access controls. Which TWO selections represent the correct investigation and remediation together? Select TWO.

  1. Inspect quota errors, retry behavior, queue occupancy, tail latency and autoscaling upper limits.
  2. Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses.
  3. Compare requested entity key bytes, stored feature-group schema, casting rules and online lookup failures.
  4. Set a justified warm minimum or scheduled capacity and balance initialization overhead against idle cost.
  5. Add guarded traffic shifts and automatic rollback to the retained healthy revision when threshold alarms trip.

Explanation

Correct Answers

A:

Inspect quota errors, retry behavior, queue occupancy, tail latency and autoscaling upper limits. This inspects evidence relevant to predictive serving rate limit and backpressure. Unlimited retries amplify overload; backpressure and controlled concurrency are needed along with scaling. After identifying the cause, the matching correction is: Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses. This matters here because during a staged restoration test at a secondary serving deployment, the following condition occurs: A public….

B:

Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses. This is the targeted corrective response for predictive serving rate limit and backpressure. Unlimited retries amplify overload; backpressure and controlled concurrency are needed along with scaling. Verify the diagnosis using: Inspect quota errors, retry behavior, queue occupancy, tail latency and autoscaling upper limits. Confirm success using: Peak traffic is rate-limited predictably and the serving fleet recovers without an unbounded request backlog. This matters here because during a staged restoration test at a secondary serving deployment, the following condition occurs: A public….

Incorrect Answers

C:

Compare requested entity key bytes, stored feature-group schema, casting rules and online lookup failures. This is an action or diagnostic relevant to feature online entity key mismatch, not one of the required two steps. In this scenario verify: Inspect quota errors, retry behavior, queue occupancy, tail latency and autoscaling upper limits. Then correct: Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses.

D:

Set a justified warm minimum or scheduled capacity and balance initialization overhead against idle cost. This is an action or diagnostic relevant to autoscaling minimum capacity versus cold load, not one of the required two steps. In this scenario verify: Inspect quota errors, retry behavior, queue occupancy, tail latency and autoscaling upper limits. Then correct: Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses.

E:

Add guarded traffic shifts and automatic rollback to the retained healthy revision when threshold alarms trip. This is an action or diagnostic relevant to model traffic rollback on high error rate, not one of the required two steps. In this scenario verify: Inspect quota errors, retry behavior, queue occupancy, tail latency and autoscaling upper limits. Then correct: Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses.

Question 1138

While reproducing a training orchestration incident in isolated staging, the following condition occurs: A CI/CD workflow always deploys the most recent checkpoint even when its validation F1 is below the previous approved baseline. The team requires an idempotent pipeline behavior and traceable evidence. Which TWO complementary choices jointly establish the evidence and appropriate correction? Select TWO.

  1. Inspect incoming schema versions, validation logs, casting rules and downstream prediction anomalies.
  2. Inspect model-metric artifacts, comparison rules, registry status and deployment dependencies.
  3. Create immutable versioned input snapshots or partition manifests and record them in training metadata.
  4. Require candidate evaluation against a governed baseline with an enforced acceptance threshold before promotion.
  5. Include immutable dataset versions and code revisions as explicit component inputs or disable inappropriate caching.

Explanation

Correct Answers

B:

Inspect model-metric artifacts, comparison rules, registry status and deployment dependencies. This inspects evidence relevant to model evaluation gate before deployment. Evaluation without a blocking release decision cannot protect production from quality regressions. After identifying the cause, the matching correction is: Require candidate evaluation against a governed baseline with an enforced acceptance threshold before promotion. This matters here because while reproducing a training orchestration incident in isolated staging, the following condition occurs: A CI/CD….

D:

Require candidate evaluation against a governed baseline with an enforced acceptance threshold before promotion. This is the targeted corrective response for model evaluation gate before deployment. Evaluation without a blocking release decision cannot protect production from quality regressions. Verify the diagnosis using: Inspect model-metric artifacts, comparison rules, registry status and deployment dependencies. Confirm success using: An intentionally weak candidate cannot deploy, whereas an approved passing model can proceed with traceable evidence. This matters here because while reproducing a training orchestration incident in isolated staging, the following condition occurs: A CI/CD….

Incorrect Answers

A:

Inspect incoming schema versions, validation logs, casting rules and downstream prediction anomalies. This is an action or diagnostic relevant to serving schema validation component, not one of the required two steps. In this scenario verify: Inspect model-metric artifacts, comparison rules, registry status and deployment dependencies. Then correct: Require candidate evaluation against a governed baseline with an enforced acceptance threshold before promotion.

C:

Create immutable versioned input snapshots or partition manifests and record them in training metadata. This is an action or diagnostic relevant to data version reproducibility in ct jobs, not one of the required two steps. In this scenario verify: Inspect model-metric artifacts, comparison rules, registry status and deployment dependencies. Then correct: Require candidate evaluation against a governed baseline with an enforced acceptance threshold before promotion.

E:

Include immutable dataset versions and code revisions as explicit component inputs or disable inappropriate caching. This is an action or diagnostic relevant to kubeflow component cache invalidation, not one of the required two steps. In this scenario verify: Inspect model-metric artifacts, comparison rules, registry status and deployment dependencies. Then correct: Require candidate evaluation against a governed baseline with an enforced acceptance threshold before promotion.

Question 1139

Within a jointly maintained Google Cloud data science workspace, the following condition occurs: A credit service looks up a feature by account number online but the training table keys records by customer identifier, causing inconsistent histories. Collaborators must avoid expanding data access or breaking experiment reproducibility. Which evidence most directly establishes what is wrong with this AI or ML workflow?

  1. Compare entity IDs, key types, online feature-view definitions, offline joins and records for the same verified customer.
  2. Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs.
  3. Inspect the labeling payload schema, data classification, annotator roles and fields needed to complete the labeling task.
  4. Inspect feature quantiles, physical validity bounds, missingness and training-only fit statistics against independent validation data.

Explanation

Correct Answer

A:

Compare entity IDs, key types, online feature-view definitions, offline joins and records for the same verified customer. This inspects evidence relevant to feature store online and offline entity alignment. Feature stores rely on consistent entity keys; incompatible identifiers lead to silent training-serving misalignment. After identifying the cause, the matching correction is: Standardize stable entity identifiers and validated feature definitions across training and online serving pathways. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: A credit service….

Incorrect Answers

B:

Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs. This examines experiment metric comparison with mismatched splits rather than the specific mechanism reported here. It is not the decisive first check for feature store online and offline entity alignment. The relevant evidence is: Compare entity IDs, key types, online feature-view definitions, offline joins and records for the same verified customer.

C:

Inspect the labeling payload schema, data classification, annotator roles and fields needed to complete the labeling task. This examines data minimization for labeling inputs rather than the specific mechanism reported here. It is not the decisive first check for feature store online and offline entity alignment. The relevant evidence is: Compare entity IDs, key types, online feature-view definitions, offline joins and records for the same verified customer.

D:

Inspect feature quantiles, physical validity bounds, missingness and training-only fit statistics against independent validation data. This examines outlier handling before feature standardization rather than the specific mechanism reported here. It is not the decisive first check for feature store online and offline entity alignment. The relevant evidence is: Compare entity IDs, key types, online feature-view definitions, offline joins and records for the same verified customer. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: A credit service….

Question 1140

As the team tests a new governed AI release against adversarial and normal data, the following condition occurs: A public conversational app relies on a single instruction asking the model to avoid prohibited content but does not test or screen generated responses. The fix must retain legitimate model usefulness and appropriate controls. Which diagnostic check should be performed FIRST to isolate the described cause?

  1. Examine explanation distributions, feature generation changes, protected attribute proxies and prediction outcomes by cohort.
  2. Review the business safety taxonomy, representative adversarial evaluation prompts, safety-filter configuration and escalation design.
  3. Inspect cohort conversion metrics, assignment validity, minimum sample requirements and the release guardrail configuration.
  4. Inspect explanation method assumptions, correlated input features, counterfactual evidence and the model's actual causal validity.

Explanation

Correct Answer

B:

Review the business safety taxonomy, representative adversarial evaluation prompts, safety-filter configuration and escalation design. This inspects evidence relevant to gen ai output safety policy enforcement. Prompt wording alone is not a reliable comprehensive safety boundary for public generative applications. After identifying the cause, the matching correction is: Apply supported safety filters and output policy checks with a controlled refusal or human review path for blocked results. This matters here because as the team tests a new governed AI release against adversarial and normal data, the following condition occurs: A….

Incorrect Answers

A:

Examine explanation distributions, feature generation changes, protected attribute proxies and prediction outcomes by cohort. This examines feature attribution drift after deployment rather than the specific mechanism reported here. It is not the decisive first check for gen ai output safety policy enforcement. The relevant evidence is: Review the business safety taxonomy, representative adversarial evaluation prompts, safety-filter configuration and escalation design.

C:

Inspect cohort conversion metrics, assignment validity, minimum sample requirements and the release guardrail configuration. This examines production incident rollback based on metrics rather than the specific mechanism reported here. It is not the decisive first check for gen ai output safety policy enforcement. The relevant evidence is: Review the business safety taxonomy, representative adversarial evaluation prompts, safety-filter configuration and escalation design.

D:

Inspect explanation method assumptions, correlated input features, counterfactual evidence and the model's actual causal validity. This examines explainability versus causal claim rather than the specific mechanism reported here. It is not the decisive first check for gen ai output safety policy enforcement. The relevant evidence is: Review the business safety taxonomy, representative adversarial evaluation prompts, safety-filter configuration and escalation design.

Professional Machine Learning Engineer Practice Tests