Google Professional Machine Learning Engineer Practice Test 77
Question 760
While reproducing a training orchestration incident in isolated staging, the following condition occurs: An orchestrated data pipeline inserts training rows twice after transient failure because a retried task appends without a stable batch identifier. The team requires an idempotent pipeline behavior and traceable evidence. Which result is the strongest postchange evidence of success?
- Running the same task twice produces one logical batch and preserves prior validated partitions.
- The deployed model and container can be traced to a particular approved commit and evaluation result.
- Changing the dataset version forces a rebuild while identical inputs may reuse validated cached results.
- An incompatible payload fails visibly while compatible older versions follow documented conversion rules.
Explanation
Correct Answer
A:Running the same task twice produces one logical batch and preserves prior validated partitions. This is specific acceptance evidence for airflow task idempotency on retry. Orchestrators often retry tasks; duplicate side effects are prevented by idempotent data operations. It confirms the controlled correction: Make the load idempotent with keyed merge or overwrite-per-version semantics and safe retry handling. This matters here because while reproducing a training orchestration incident in isolated staging, the following condition occurs: An….
Incorrect Answers
B:The deployed model and container can be traced to a particular approved commit and evaluation result. This result could demonstrate success for cloud build release version promotion but not for the stated airflow task idempotency on retry issue. The direct test is: Running the same task twice produces one logical batch and preserves prior validated partitions.
C:Changing the dataset version forces a rebuild while identical inputs may reuse validated cached results. This result could demonstrate success for kubeflow component cache invalidation but not for the stated airflow task idempotency on retry issue. The direct test is: Running the same task twice produces one logical batch and preserves prior validated partitions.
D:An incompatible payload fails visibly while compatible older versions follow documented conversion rules. This result could demonstrate success for serving schema validation component but not for the stated airflow task idempotency on retry issue. The direct test is: Running the same task twice produces one logical batch and preserves prior validated partitions.
Question 761
During a postdeployment AI safety and reliability review, the following condition occurs: A company assistant follows instructions embedded in an untrusted retrieved webpage, overriding the system-approved task and attempting to reveal internal data. The model owner must provide repeatable evidence rather than a cosmetic dashboard change. Which outcome most directly demonstrates a successful, measurable fix?
- Attack-bearing documents fail to redirect privileged behavior while legitimate grounded answers remain available.
- Test prompts preserve support utility without exposing real account identifiers to unauthorized model requests.
- Red-team test prompts trigger expected safeguards while legitimate requests continue to receive useful responses.
- A synthetic distribution shift triggers an attributable alert while an unchanged control sample stays within tolerance.
Explanation
Correct Answer
A:Attack-bearing documents fail to redirect privileged behavior while legitimate grounded answers remain available. This is specific acceptance evidence for prompt injection through retrieved documents. External retrieval content can carry attacker instructions; data should not receive the authority of trusted application instructions. It confirms the controlled correction: Treat retrieved text as untrusted data, isolate instructions, restrict tool actions and apply appropriate Model Armor screening.
Incorrect Answers
B:Test prompts preserve support utility without exposing real account identifiers to unauthorized model requests. This result could demonstrate success for sensitive identifiers leaked to llm prompts but not for the stated prompt injection through retrieved documents issue. The direct test is: Attack-bearing documents fail to redirect privileged behavior while legitimate grounded answers remain available.
C:Red-team test prompts trigger expected safeguards while legitimate requests continue to receive useful responses. This result could demonstrate success for gen ai output safety policy enforcement but not for the stated prompt injection through retrieved documents issue. The direct test is: Attack-bearing documents fail to redirect privileged behavior while legitimate grounded answers remain available.
D:A synthetic distribution shift triggers an attributable alert while an unchanged control sample stays within tolerance. This result could demonstrate success for input data distribution drift alarm but not for the stated prompt injection through retrieved documents issue. The direct test is: Attack-bearing documents fail to redirect privileged behavior while legitimate grounded answers remain available.
Question 762
While reproducing a training orchestration incident in isolated staging, the following condition occurs: An operations team retrains its classifier every hour based on a clock even when data have not changed, wasting compute and risking unreviewed model swaps. The team requires an idempotent pipeline behavior and traceable evidence. Which evidence most directly establishes what is wrong with this AI or ML workflow?
- Inspect Cloud Build substitutions, commit SHAs, model registry versions, signed image digests and approval records.
- Inspect snapshot identifiers, SQL filters, table partitions and the model version's recorded input manifest.
- Inspect component input parameters, artifact provenance, caching settings and source snapshot identifiers.
- Inspect training data freshness, drift signals, label availability, business cadence and automated promotion rules.
Explanation
Correct Answer
D:Inspect training data freshness, drift signals, label availability, business cadence and automated promotion rules. This inspects evidence relevant to data drift triggered retraining policy. Retraining on a timer alone does not imply better models and may use incomplete or unrepresentative labels. After identifying the cause, the matching correction is: Trigger candidate retraining only when justified evidence and sufficient labeled data are available, with evaluation gates. This matters here because while reproducing a training orchestration incident in isolated staging, the following condition occurs: An….
Incorrect Answers
A:Inspect Cloud Build substitutions, commit SHAs, model registry versions, signed image digests and approval records. This examines cloud build release version promotion rather than the specific mechanism reported here. It is not the decisive first check for data drift triggered retraining policy. The relevant evidence is: Inspect training data freshness, drift signals, label availability, business cadence and automated promotion rules.
B:Inspect snapshot identifiers, SQL filters, table partitions and the model version's recorded input manifest. This examines data version reproducibility in ct jobs rather than the specific mechanism reported here. It is not the decisive first check for data drift triggered retraining policy. The relevant evidence is: Inspect training data freshness, drift signals, label availability, business cadence and automated promotion rules. This matters here because while reproducing a training orchestration incident in isolated staging, the following condition occurs: An….
C:Inspect component input parameters, artifact provenance, caching settings and source snapshot identifiers. This examines kubeflow component cache invalidation rather than the specific mechanism reported here. It is not the decisive first check for data drift triggered retraining policy. The relevant evidence is: Inspect training data freshness, drift signals, label availability, business cadence and automated promotion rules.
Question 763
While preparing a production training workflow on Gemini Enterprise Agent Platform, the following condition occurs: A fraud team with 40 engineered numeric transaction features starts with a costly vision transformer while ignoring a strong tabular baseline. Engineers need an auditable, resource-aware correction to the issue. Which outcome most directly demonstrates a successful, measurable fix?
- Every evaluation feature is available at its simulated prediction time and later-period scores remain stable.
- Relevant articles rank highly for unseen paraphrased queries and retrieval hit rate is measured against judgments.
- The chosen threshold detects clinically relevant positives while the false-alarm rate is documented for deployment review.
- The selected candidate meets fraud recall, inference latency and cost limits on a representative untouched split.
Explanation
Correct Answer
D:The selected candidate meets fraud recall, inference latency and cost limits on a representative untouched split. This is specific acceptance evidence for deep model versus gradient boosted trees on tabular. Structured tabular datasets often have competitive simple baselines; architectural complexity is not inherently better. It confirms the controlled correction: Benchmark a suitable boosted-tree model against more complex candidates and retain complexity only when justified by held-out gains. This matters here because while preparing a production training workflow on Gemini Enterprise Agent Platform, the following condition occurs:….
Incorrect Answers
A:Every evaluation feature is available at its simulated prediction time and later-period scores remain stable. This result could demonstrate success for temporal training label leakage in rolling windows but not for the stated deep model versus gradient boosted trees on tabular issue. The direct test is: The selected candidate meets fraud recall, inference latency and cost limits on a representative untouched split.
B:Relevant articles rank highly for unseen paraphrased queries and retrieval hit rate is measured against judgments. This result could demonstrate success for embedding retrieval versus text classifier but not for the stated deep model versus gradient boosted trees on tabular issue. The direct test is: The selected candidate meets fraud recall, inference latency and cost limits on a representative untouched split. This matters here because while preparing a production training workflow on Gemini Enterprise Agent Platform, the following condition occurs:….
C:The chosen threshold detects clinically relevant positives while the false-alarm rate is documented for deployment review. This result could demonstrate success for cost-sensitive threshold for rare positives but not for the stated deep model versus gradient boosted trees on tabular issue. The direct test is: The selected candidate meets fraud recall, inference latency and cost limits on a representative untouched split.
Question 764
As the team tests a new governed AI release against adversarial and normal data, the following condition occurs: A transaction fraud model's feature distributions stay nearly unchanged while the relationship between inputs and observed fraud outcomes deteriorates. The fix must retain legitimate model usefulness and appropriate controls. Which statement identifies the primary reason for this failure?
- Excess prompt context can increase processing cost and latency without improving response accuracy.
- Feature attribution drift can indicate changing model reliance even when overall score distributions remain stable.
- HTTP health is not evidence that incoming feature distributions still resemble training data.
- Concept drift changes conditional label relationships and may occur without obvious marginal feature drift.
Explanation
Correct Answer
D:Concept drift changes conditional label relationships and may occur without obvious marginal feature drift. This explains the mechanism behind concept drift with stable feature statistics and why the described failure is possible. A discriminating investigation is: Compare time-stratified labeled performance, feature distribution stability and delayed ground truth coverage. The corresponding repair is: Monitor outcome-based quality metrics on matured cohorts and retrain when verified concept drift is significant.
Incorrect Answers
A:Excess prompt context can increase processing cost and latency without improving response accuracy. This principle describes llm latency and cost token regression, which is not the cause of the observed concept drift with stable feature statistics failure. The controlling explanation is: Concept drift changes conditional label relationships and may occur without obvious marginal feature drift.
B:Feature attribution drift can indicate changing model reliance even when overall score distributions remain stable. This principle describes feature attribution drift after deployment, which is not the cause of the observed concept drift with stable feature statistics failure. The controlling explanation is: Concept drift changes conditional label relationships and may occur without obvious marginal feature drift.
C:HTTP health is not evidence that incoming feature distributions still resemble training data. This principle describes input data distribution drift alarm, which is not the cause of the observed concept drift with stable feature statistics failure. The controlling explanation is: Concept drift changes conditional label relationships and may occur without obvious marginal feature drift.
Question 765
Within a jointly maintained Google Cloud data science workspace, the following condition occurs: A distributed preprocessing job joins two huge feature tables, but one NULL-like customer category contains most rows and one reducer becomes the bottleneck. Collaborators must avoid expanding data access or breaking experiment reproducibility. Which diagnostic check should be performed FIRST to isolate the described cause?
- Examine the partitioning column, query plan, bytes scanned and predicate form against the table's partition definition.
- Inspect feature event timestamps, prediction cutoffs, feature joins and any post-cancellation administrative flags.
- Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions.
- Inspect join-key frequency distributions, per-task runtime, shuffle spill and the output cardinality before optimization.
Explanation
Correct Answer
D:Inspect join-key frequency distributions, per-task runtime, shuffle spill and the output cardinality before optimization. This inspects evidence relevant to managed spark join-key skew repair. Skew concentrates otherwise distributed work on a single hot key, which extra workers alone may not resolve. After identifying the cause, the matching correction is: Handle the dominant key separately or repartition and salt the skewed join while validating join semantics and counts. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: A distributed….
Incorrect Answers
A:Examine the partitioning column, query plan, bytes scanned and predicate form against the table's partition definition. This examines bigquery partition filter cost optimization rather than the specific mechanism reported here. It is not the decisive first check for managed spark join-key skew repair. The relevant evidence is: Inspect join-key frequency distributions, per-task runtime, shuffle spill and the output cardinality before optimization. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: A distributed….
B:Inspect feature event timestamps, prediction cutoffs, feature joins and any post-cancellation administrative flags. This examines point-in-time feature leakage prevention rather than the specific mechanism reported here. It is not the decisive first check for managed spark join-key skew repair. The relevant evidence is: Inspect join-key frequency distributions, per-task runtime, shuffle spill and the output cardinality before optimization.
C:Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions. This examines ml metadata lineage for production model rather than the specific mechanism reported here. It is not the decisive first check for managed spark join-key skew repair. The relevant evidence is: Inspect join-key frequency distributions, per-task runtime, shuffle spill and the output cardinality before optimization.
Question 766
At a Google Cloud analytics team preparing a governed pilot, the following condition occurs: A retailer has no trustworthy segment labels but attempts a supervised classifier with synthetic classes supplied by the sales team. The team must satisfy a reproducible low-code AI assessment. Select TWO: the decisive diagnostic step and the targeted repair for this specific issue.
- Exclude post-outcome fields, create point-in-time valid training and evaluation partitions, and rerun ML.EVALUATE.
- Inspect the timestamp ordering, seasonal patterns, horizon requirement, and the BigQuery ML model_type and time-series options.
- Benchmark a supported lower-latency Gemini variant against the current model under representative traffic before routing requests.
- Check whether real class labels exist and examine the standardized input distributions and cluster stability under different seeds.
- Train an appropriately scaled BigQuery ML KMEANS model and assess cluster cohesion, size and business interpretability.
Explanation
Correct Answers
D:Check whether real class labels exist and examine the standardized input distributions and cluster stability under different seeds. This inspects evidence relevant to bigquery ml kmeans customer segmentation. K-means discovers unlabeled similarity groups; fabricating labels would make supervised accuracy meaningless. After identifying the cause, the matching correction is: Train an appropriately scaled BigQuery ML KMEANS model and assess cluster cohesion, size and business interpretability. Here, the relevant situation is: the following condition occurs: A retailer has no trustworthy segment labels but attempts a supervised classifier….
E:Train an appropriately scaled BigQuery ML KMEANS model and assess cluster cohesion, size and business interpretability. This is the targeted corrective response for bigquery ml kmeans customer segmentation. K-means discovers unlabeled similarity groups; fabricating labels would make supervised accuracy meaningless. Verify the diagnosis using: Check whether real class labels exist and examine the standardized input distributions and cluster stability under different seeds. Confirm success using: Stable clusters reveal coherent customer behavior without pretending an unsupervised assignment is known supervised ground truth. Here, the relevant situation is: the following condition occurs: A retailer has no trustworthy segment labels but attempts a supervised classifier….
Incorrect Answers
A:Exclude post-outcome fields, create point-in-time valid training and evaluation partitions, and rerun ML.EVALUATE. This is an action or diagnostic relevant to bigquery ml ml.evaluate leakage detection, not one of the required two steps. In this scenario verify: Check whether real class labels exist and examine the standardized input distributions and cluster stability under different seeds. Then correct: Train an appropriately scaled BigQuery ML KMEANS model and assess cluster cohesion, size and business interpretability.
B:Inspect the timestamp ordering, seasonal patterns, horizon requirement, and the BigQuery ML model_type and time-series options. This is an action or diagnostic relevant to bigquery arima_plus forecast horizon, not one of the required two steps. In this scenario verify: Check whether real class labels exist and examine the standardized input distributions and cluster stability under different seeds. Then correct: Train an appropriately scaled BigQuery ML KMEANS model and assess cluster cohesion, size and business interpretability.
C:Benchmark a supported lower-latency Gemini variant against the current model under representative traffic before routing requests. This is an action or diagnostic relevant to gemini model choice for latency versus complexity, not one of the required two steps. In this scenario verify: Check whether real class labels exist and examine the standardized input distributions and cluster stability under different seeds. Then correct: Train an appropriately scaled BigQuery ML KMEANS model and assess cluster cohesion, size and business interpretability.
Question 767
During a production ML inference release on Google Cloud, the following condition occurs: A real-time scoring service reads a managed feature view that has not synchronized since yesterday while the source BigQuery table contains recent transaction updates. The service must meet its agreed quality and latency requirements. Which underlying engineering principle explains why the original approach fails?
- Online feature serving requires an up-to-date synchronized view; source-table changes need not appear instantly.
- Network distance can dominate inference response times when the actual model execution is already short.
- Lower precision can improve serving efficiency but may harm numerical accuracy, so it must be validated.
- Embedding matrices interpret token IDs according to one vocabulary; mismatched preprocessing changes input meaning.
Explanation
Correct Answer
A:Online feature serving requires an up-to-date synchronized view; source-table changes need not appear instantly. This explains the mechanism behind feature store freshness and online sync and why the described failure is possible. A discriminating investigation is: Inspect feature-view sync status, latest source timestamps, entity ID and online lookup freshness. The corresponding repair is: Schedule and monitor approved Feature Store synchronization at the needed frequency, with stale-feature handling. This matters here because a real-time scoring service reads a managed feature view that has not synchronized since yesterday while the source….
Incorrect Answers
B:Network distance can dominate inference response times when the actual model execution is already short. This principle describes serving region proximity versus model compute, which is not the cause of the observed feature store freshness and online sync failure. The controlling explanation is: Online feature serving requires an up-to-date synchronized view; source-table changes need not appear instantly.
C:Lower precision can improve serving efficiency but may harm numerical accuracy, so it must be validated. This principle describes inference optimization using quantization, which is not the cause of the observed feature store freshness and online sync failure. The controlling explanation is: Online feature serving requires an up-to-date synchronized view; source-table changes need not appear instantly.
D:Embedding matrices interpret token IDs according to one vocabulary; mismatched preprocessing changes input meaning. This principle describes inference preprocessing tokenization mismatch, which is not the cause of the observed feature store freshness and online sync failure. The controlling explanation is: Online feature serving requires an up-to-date synchronized view; source-table changes need not appear instantly.
Question 768
During a regulated MLOps pipeline deployment, the following condition occurs: A feature-building pipeline reuses a cached output after source data changed because the component's cache key does not incorporate dataset version. The workflow must be repeatable and protect the last approved production model. Which technical explanation best accounts for the reported behavior?
- Orchestration needs explicit data and control dependencies; notebook execution order is not a production DAG.
- Model retraining reproducibility requires stable inputs rather than unversioned reads of an evolving table.
- Orchestrators often retry tasks; duplicate side effects are prevented by idempotent data operations.
- Pipeline caches assume identical declared inputs; hidden external dependencies can make cached outputs stale.
Explanation
Correct Answer
D:Pipeline caches assume identical declared inputs; hidden external dependencies can make cached outputs stale. This explains the mechanism behind kubeflow component cache invalidation and why the described failure is possible. A discriminating investigation is: Inspect component input parameters, artifact provenance, caching settings and source snapshot identifiers. The corresponding repair is: Include immutable dataset versions and code revisions as explicit component inputs or disable inappropriate caching.
Incorrect Answers
A:Orchestration needs explicit data and control dependencies; notebook execution order is not a production DAG. This principle describes pipeline component dependency order, which is not the cause of the observed kubeflow component cache invalidation failure. The controlling explanation is: Pipeline caches assume identical declared inputs; hidden external dependencies can make cached outputs stale.
B:Model retraining reproducibility requires stable inputs rather than unversioned reads of an evolving table. This principle describes data version reproducibility in ct jobs, which is not the cause of the observed kubeflow component cache invalidation failure. The controlling explanation is: Pipeline caches assume identical declared inputs; hidden external dependencies can make cached outputs stale.
C:Orchestrators often retry tasks; duplicate side effects are prevented by idempotent data operations. This principle describes airflow task idempotency on retry, which is not the cause of the observed kubeflow component cache invalidation failure. The controlling explanation is: Pipeline caches assume identical declared inputs; hidden external dependencies can make cached outputs stale.
Question 769
Within a jointly maintained Google Cloud data science workspace, the following condition occurs: An auditor cannot identify which training dataset and code commit produced a deployed model because the model archive has only a human-readable name. Collaborators must avoid expanding data access or breaking experiment reproducibility. Which evidence most directly establishes what is wrong with this AI or ML workflow?
- Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions.
- Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs.
- Inspect judge rubric prompts, evaluation dataset balance, repeated ratings, human benchmark and uncertainty in pairwise preferences.
- Inspect saved artifact format, framework versions, device placement requirements and the model card's supported execution paths.
Explanation
Correct Answer
A:Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions. This inspects evidence relevant to ml metadata lineage for production model. MLOps provenance requires a recorded chain from code and data to model and deployment, not merely a friendly name. After identifying the cause, the matching correction is: Record input dataset versions, training runs, code revisions and output artifact links in the managed metadata and registry workflow. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: An auditor cannot….
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 ml metadata lineage for production model. The relevant evidence is: Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions.
C:Inspect judge rubric prompts, evaluation dataset balance, repeated ratings, human benchmark and uncertainty in pairwise preferences. This examines llm-as-a-judge rubric reliability rather than the specific mechanism reported here. It is not the decisive first check for ml metadata lineage for production model. The relevant evidence is: Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: An auditor cannot….
D:Inspect saved artifact format, framework versions, device placement requirements and the model card's supported execution paths. This examines model garden jax accelerator compatibility rather than the specific mechanism reported here. It is not the decisive first check for ml metadata lineage for production model. The relevant evidence is: Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions.
Question 770
At a Google Cloud analytics team preparing a governed pilot, the following condition occurs: A support assistant includes a full 200-page manual in every prompt despite users asking brief questions, producing excessive latency and cost. The team must satisfy a reproducible low-code AI assessment. Which configuration or procedure is the most appropriate correction?
- Select a permitted task-compatible Gemini or other supported foundation model and benchmark it against simpler candidates.
- Retrieve only relevant passages with provenance and use a concise structured prompt with measured token limits.
- Use a BigQuery ML LINEAR_REG or suitable regression model with a numeric target and evaluate continuous-error metrics on held-out records.
- Use Cloud Translation API for supported text translation, then separately validate downstream classification against localized content.
Explanation
Correct Answer
B:Retrieve only relevant passages with provenance and use a concise structured prompt with measured token limits. This is the targeted corrective response for gemini prompt context token budget. Unfocused context consumes tokens and can distract generation; targeted retrieval can reduce cost without discarding needed facts. Verify the diagnosis using: Inspect prompt token usage, retrieval relevance, model context requirements, answer grounding and p95 latency across queries. Confirm success using: On held-out questions, answers remain grounded while input tokens and p95 response latency decline.
Incorrect Answers
A:Select a permitted task-compatible Gemini or other supported foundation model and benchmark it against simpler candidates. This action addresses model garden task and license selection, not the failure in the described gemini prompt context token budget case. The targeted correction is: Retrieve only relevant passages with provenance and use a concise structured prompt with measured token limits.
C:Use a BigQuery ML LINEAR_REG or suitable regression model with a numeric target and evaluate continuous-error metrics on held-out records. This action addresses bigquery ml linear regression for continuous value, not the failure in the described gemini prompt context token budget case. The targeted correction is: Retrieve only relevant passages with provenance and use a concise structured prompt with measured token limits.
D:Use Cloud Translation API for supported text translation, then separately validate downstream classification against localized content. This action addresses translate api for multilingual product descriptions, not the failure in the described gemini prompt context token budget case. The targeted correction is: Retrieve only relevant passages with provenance and use a concise structured prompt with measured token limits.