Google Professional Machine Learning Engineer Practice Test 43
Question 420
During a staged restoration test at a secondary serving deployment, the following condition occurs: A multi-class model returns numeric class indices, but an API gateway uses a stale alphabetical mapping and mislabels predictions. The team must prove an actual user-visible improvement without weakening access controls. Which result is the strongest postchange evidence of success?
- Identical sample text yields matching token IDs and model outputs in offline and online environments.
- Identical release manifests resolve to identical image bytes across reproducible deployments.
- Every controlled sample's returned label matches the model's trained class index with no misordered classes.
- Previously missed customer IDs retrieve the intended feature values without collapsing distinct leading-zero identifiers.
Explanation
Correct Answer
C:Every controlled sample's returned label matches the model's trained class index with no misordered classes. This is specific acceptance evidence for custom postprocessing label mapping drift. Index ordering is part of the model contract; separate stale label dictionaries can silently corrupt inference. It confirms the controlled correction: Ship the class vocabulary with the model artifact and version or validate the output adapter against it.
Incorrect Answers
A:Identical sample text yields matching token IDs and model outputs in offline and online environments. This result could demonstrate success for inference preprocessing tokenization mismatch but not for the stated custom postprocessing label mapping drift issue. The direct test is: Every controlled sample's returned label matches the model's trained class index with no misordered classes.
B:Identical release manifests resolve to identical image bytes across reproducible deployments. This result could demonstrate success for model artifact image digest reproducibility but not for the stated custom postprocessing label mapping drift issue. The direct test is: Every controlled sample's returned label matches the model's trained class index with no misordered classes.
D:Previously missed customer IDs retrieve the intended feature values without collapsing distinct leading-zero identifiers. This result could demonstrate success for feature online entity key mismatch but not for the stated custom postprocessing label mapping drift issue. The direct test is: Every controlled sample's returned label matches the model's trained class index with no misordered classes.
Question 421
While reproducing a training orchestration incident in isolated staging, the following condition occurs: A build pushes a model-serving image into production even after its dependency scan reports a critical vulnerability. The team requires an idempotent pipeline behavior and traceable evidence. Which targeted engineering change most directly addresses the failure?
- Tune component resource requests/limits and stream or partition embeddings rather than materializing every vector at once.
- Enforce a shared typed schema contract and reject incompatible feature versions before training or inference.
- Use a retraining policy with significant trigger thresholds, independent evaluation and staged deployment approval.
- Fail protected build or release stages when prohibited vulnerabilities exist and require documented exceptions for approved deviations.
Explanation
Correct Answer
D:Fail protected build or release stages when prohibited vulnerabilities exist and require documented exceptions for approved deviations. This is the targeted corrective response for cloud build container security gate. Security scanning provides protection only if the CI/CD pipeline enforces its result before release. Verify the diagnosis using: Inspect Cloud Build stage status propagation, vulnerability findings, policy thresholds and deployment triggers. Confirm success using: A deliberately vulnerable test image is blocked from production while clean images pass the same gate.
Incorrect Answers
A:Tune component resource requests/limits and stream or partition embeddings rather than materializing every vector at once. This action addresses custom kfp component resource limits, not the failure in the described cloud build container security gate case. The targeted correction is: Fail protected build or release stages when prohibited vulnerabilities exist and require documented exceptions for approved deviations.
B:Enforce a shared typed schema contract and reject incompatible feature versions before training or inference. This action addresses serving schema validation component, not the failure in the described cloud build container security gate case. The targeted correction is: Fail protected build or release stages when prohibited vulnerabilities exist and require documented exceptions for approved deviations.
C:Use a retraining policy with significant trigger thresholds, independent evaluation and staged deployment approval. This action addresses canary failure feedback into ct policy, not the failure in the described cloud build container security gate case. The targeted correction is: Fail protected build or release stages when prohibited vulnerabilities exist and require documented exceptions for approved deviations.
Question 422
Within a jointly maintained Google Cloud data science workspace, the following condition occurs: A prototype notebook commits a service-account JSON private key beside its feature code so colleagues can open Cloud Storage directly. Collaborators must avoid expanding data access or breaking experiment reproducibility. Which diagnostic check should be performed FIRST to isolate the described cause?
- Restart the runtime, run the notebook top to bottom and compare variable initialization with committed source revision.
- Inspect saved artifact format, framework versions, device placement requirements and the model card's supported execution paths.
- Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs.
- Inspect repository history, mounted credentials, authorized identities and any logs of secret access from the shared project.
Explanation
Correct Answer
D:Inspect repository history, mounted credentials, authorized identities and any logs of secret access from the shared project. This inspects evidence relevant to workbench credentials embedded in repository. Long-lived secrets copied into notebooks or Git can be reused outside intended workloads and must be rotated. After identifying the cause, the matching correction is: Revoke and rotate the exposed key and use managed workload identity or approved short-lived credentials with least privilege.
Incorrect Answers
A:Restart the runtime, run the notebook top to bottom and compare variable initialization with committed source revision. This examines notebook transient state and execution order rather than the specific mechanism reported here. It is not the decisive first check for workbench credentials embedded in repository. The relevant evidence is: Inspect repository history, mounted credentials, authorized identities and any logs of secret access from the shared project.
B: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 workbench credentials embedded in repository. The relevant evidence is: Inspect repository history, mounted credentials, authorized identities and any logs of secret access from the shared project.
C: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 workbench credentials embedded in repository. The relevant evidence is: Inspect repository history, mounted credentials, authorized identities and any logs of secret access from the shared project.
Question 423
During a regulated MLOps pipeline deployment, the following condition occurs: A build pipeline deploys an arbitrary model tag without tying it to the source revision and the approved evaluation report. The workflow must be repeatable and protect the last approved production model. Which result is the strongest postchange evidence of success?
- No-op periods create no unnecessary models and validated drift events produce candidates requiring approval.
- A previous model can be rebuilt from its exact documented training dataset and code revision.
- Representative raw records produce equal feature vectors offline and online, including unseen-category behavior.
- The deployed model and container can be traced to a particular approved commit and evaluation result.
Explanation
Correct Answer
D:The deployed model and container can be traced to a particular approved commit and evaluation result. This is specific acceptance evidence for cloud build release version promotion. Unpinned tags do not establish a trustworthy relationship between code, model artifacts and production rollout. It confirms the controlled correction: Promote explicit immutable model and image versions with versioned release manifests and review gates.
Incorrect Answers
A: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 cloud build release version promotion issue. The direct test is: The deployed model and container can be traced to a particular approved commit and evaluation result.
B:A previous model can be rebuilt from its exact documented training dataset and code revision. This result could demonstrate success for data version reproducibility in ct jobs but not for the stated cloud build release version promotion issue. The direct test is: The deployed model and container can be traced to a particular approved commit and evaluation result.
C:Representative raw records produce equal feature vectors offline and online, including unseen-category behavior. This result could demonstrate success for training serving preprocessing artifact sharing but not for the stated cloud build release version promotion issue. The direct test is: The deployed model and container can be traced to a particular approved commit and evaluation result.
Question 424
During a controlled replay of the data and model handoff to a second engineering team, the following condition occurs: An image labeling task sends full customer identity and contact records to external annotators although only product photographs need annotation. The handoff must preserve privacy, lineage and testable behavior. Which underlying engineering principle explains why the original approach fails?
- Choosing a managed SQL engine for warehouse-resident batch joins avoids unnecessary always-on stream processing.
- A scaler learned from corrupted or held-out values can distort typical feature magnitudes and leak evaluation information.
- Protecting data in transit does not excuse sending unnecessary sensitive attributes to a third-party workflow.
- A single stochastic training run is not sufficient to distinguish algorithm improvement from sampling variation.
Explanation
Correct Answer
C:Protecting data in transit does not excuse sending unnecessary sensitive attributes to a third-party workflow. This explains the mechanism behind data minimization for labeling inputs and why the described failure is possible. A discriminating investigation is: Inspect the labeling payload schema, data classification, annotator roles and fields needed to complete the labeling task. The corresponding repair is: Redact or tokenize unnecessary identifying fields before labeling and limit access to approved mappings. This matters here because during a controlled replay of the data and model handoff to a second engineering team, the following condition….
Incorrect Answers
A:Choosing a managed SQL engine for warehouse-resident batch joins avoids unnecessary always-on stream processing. This principle describes bigquery versus dataflow preparation selection, which is not the cause of the observed data minimization for labeling inputs failure. The controlling explanation is: Protecting data in transit does not excuse sending unnecessary sensitive attributes to a third-party workflow.
B:A scaler learned from corrupted or held-out values can distort typical feature magnitudes and leak evaluation information. This principle describes outlier handling before feature standardization, which is not the cause of the observed data minimization for labeling inputs failure. The controlling explanation is: Protecting data in transit does not excuse sending unnecessary sensitive attributes to a third-party workflow. This matters here because during a controlled replay of the data and model handoff to a second engineering team, the following condition….
D:A single stochastic training run is not sufficient to distinguish algorithm improvement from sampling variation. This principle describes experiment random seed and data order variance, which is not the cause of the observed data minimization for labeling inputs failure. The controlling explanation is: Protecting data in transit does not excuse sending unnecessary sensitive attributes to a third-party workflow.
Question 425
During a controlled replay of the data and model handoff to a second engineering team, the following condition occurs: A project declares model B superior even though A was evaluated on last quarter and B on a deliberately easier random sample. The handoff must preserve privacy, lineage and testable behavior. Select TWO: the most relevant root-cause check and the specific observation that proves the fix worked.
- 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.
- Repeated controlled runs show bounded expected variance and any real improvement exceeds run-to-run noise.
- Long-tail shuffle tasks shorten and the normalized output row count matches a reference implementation.
- Candidate scores can be reproduced on the same held-out data and their differences are attributable to model changes.
Explanation
Correct Answers
B:Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs. This inspects evidence relevant to experiment metric comparison with mismatched splits. Metrics from different data distributions are not directly comparable evidence of model improvement. After identifying the cause, the matching correction is: Register a common versioned holdout and comparable metrics in Agent Platform Experiments before ranking candidates.
E:Candidate scores can be reproduced on the same held-out data and their differences are attributable to model changes. This is specific acceptance evidence for experiment metric comparison with mismatched splits. Metrics from different data distributions are not directly comparable evidence of model improvement. It confirms the controlled correction: Register a common versioned holdout and comparable metrics in Agent Platform Experiments before ranking candidates.
Incorrect Answers
A:Inspect Agent Platform ML Metadata artifacts, pipeline executions, dataset fingerprints and deployment revisions. This relates to ml metadata lineage for production model but does not establish the requested before-and-after evidence. The actual diagnostic is: Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs. The specific acceptance result is: Candidate scores can be reproduced on the same held-out data and their differences are attributable to model changes.
C:Repeated controlled runs show bounded expected variance and any real improvement exceeds run-to-run noise. This relates to experiment random seed and data order variance but does not establish the requested before-and-after evidence. The actual diagnostic is: Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs. The specific acceptance result is: Candidate scores can be reproduced on the same held-out data and their differences are attributable to model changes.
D:Long-tail shuffle tasks shorten and the normalized output row count matches a reference implementation. This relates to managed spark join-key skew repair but does not establish the requested before-and-after evidence. The actual diagnostic is: Inspect dataset and split fingerprints, label definitions, metric versions and evaluation timestamps for both experiment runs. The specific acceptance result is: Candidate scores can be reproduced on the same held-out data and their differences are attributable to model changes.
Question 426
At a Google Cloud analytics team preparing a governed pilot, the following condition occurs: A content moderation team needs generic objects detected in product photographs but starts training a custom network without checking pretrained image analysis. The team must satisfy a reproducible low-code AI assessment. Which underlying engineering principle explains why the original approach fails?
- Object or image classification only identifies image categories; document understanding targets structured text fields.
- Managed AutoML can train supported supervised tabular tasks without requiring custom neural-network code.
- Model size alone does not ensure suitable task support, usage rights or production economics.
- A pretrained vision API may cover general image concepts, while specialized categories may need a custom model.
Explanation
Correct Answer
D:A pretrained vision API may cover general image concepts, while specialized categories may need a custom model. This explains the mechanism behind vision api pretrained image labels versus custom detection and why the described failure is possible. A discriminating investigation is: Inspect requested labels or bounding boxes, sample image quality, supported Vision API capabilities and required detection thresholds. The corresponding repair is: Evaluate Cloud Vision API for supported generic image detection and use custom training only for missing domain-specific classes.
Incorrect Answers
A:Object or image classification only identifies image categories; document understanding targets structured text fields. This principle describes document ai structured invoice extraction, which is not the cause of the observed vision api pretrained image labels versus custom detection failure. The controlling explanation is: A pretrained vision API may cover general image concepts, while specialized categories may need a custom model.
B:Managed AutoML can train supported supervised tabular tasks without requiring custom neural-network code. This principle describes automl tabular supervised model selection, which is not the cause of the observed vision api pretrained image labels versus custom detection failure. The controlling explanation is: A pretrained vision API may cover general image concepts, while specialized categories may need a custom model.
C:Model size alone does not ensure suitable task support, usage rights or production economics. This principle describes model garden task and license selection, which is not the cause of the observed vision api pretrained image labels versus custom detection failure. The controlling explanation is: A pretrained vision API may cover general image concepts, while specialized categories may need a custom model.
Question 427
Within a jointly maintained Google Cloud data science workspace, the following condition occurs: Two experiments use the same model configuration but show different accuracy due to untracked data shuffling and nondeterministic seeds. Collaborators must avoid expanding data access or breaking experiment reproducibility. Which TWO complementary choices jointly establish the evidence and appropriate correction? Select TWO.
- Inspect pipeline execution IDs, experiment tracking records, artifact URIs and the link between metrics and saved models.
- Log reproducibility settings and rerun comparable trials with fixed split and controlled randomness before attributing changes.
- Inspect random seeds, sample ordering, framework determinism settings and environment versions recorded per run.
- Push aggregations into BigQuery or stream bounded samples to the notebook rather than materializing the whole table.
- Register a common versioned holdout and comparable metrics in Agent Platform Experiments before ranking candidates.
Explanation
Correct Answers
B:Log reproducibility settings and rerun comparable trials with fixed split and controlled randomness before attributing changes. This is the targeted corrective response for experiment random seed and data order variance. A single stochastic training run is not sufficient to distinguish algorithm improvement from sampling variation. Verify the diagnosis using: Inspect random seeds, sample ordering, framework determinism settings and environment versions recorded per run. Confirm success using: Repeated controlled runs show bounded expected variance and any real improvement exceeds run-to-run noise. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: Two experiments use….
C:Inspect random seeds, sample ordering, framework determinism settings and environment versions recorded per run. This inspects evidence relevant to experiment random seed and data order variance. A single stochastic training run is not sufficient to distinguish algorithm improvement from sampling variation. After identifying the cause, the matching correction is: Log reproducibility settings and rerun comparable trials with fixed split and controlled randomness before attributing changes. This matters here because within a jointly maintained Google Cloud data science workspace, the following condition occurs: Two experiments use….
Incorrect Answers
A:Inspect pipeline execution IDs, experiment tracking records, artifact URIs and the link between metrics and saved models. This is an action or diagnostic relevant to pipeline experiment versus notebook-only tracking, not one of the required two steps. In this scenario verify: Inspect random seeds, sample ordering, framework determinism settings and environment versions recorded per run. Then correct: Log reproducibility settings and rerun comparable trials with fixed split and controlled randomness before attributing changes.
D:Push aggregations into BigQuery or stream bounded samples to the notebook rather than materializing the whole table. This is an action or diagnostic relevant to notebook memory exhaustion on full data extraction, not one of the required two steps. In this scenario verify: Inspect random seeds, sample ordering, framework determinism settings and environment versions recorded per run. Then correct: Log reproducibility settings and rerun comparable trials with fixed split and controlled randomness before attributing changes.
E:Register a common versioned holdout and comparable metrics in Agent Platform Experiments before ranking candidates. This is an action or diagnostic relevant to experiment metric comparison with mismatched splits, not one of the required two steps. In this scenario verify: Inspect random seeds, sample ordering, framework determinism settings and environment versions recorded per run. Then correct: Log reproducibility settings and rerun comparable trials with fixed split and controlled randomness before attributing changes.
Question 428
While reproducing a training orchestration incident in isolated staging, the following condition occurs: A model registry entry lacks the upstream version of the training feature table, preventing investigators from reconstructing a reported prediction. 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 component definitions, pipeline package versions, parameters and source revisions used in both environments.
- Inspect ML Metadata execution links, table snapshots, transformations and model version lineage.
- Inspect object prefixes, event filters, completed-batch manifests and recent pipeline execution IDs.
- Inspect incoming schema versions, validation logs, casting rules and downstream prediction anomalies.
Explanation
Correct Answer
B:Inspect ML Metadata execution links, table snapshots, transformations and model version lineage. This inspects evidence relevant to model feature lineage for audit. Model weights alone do not capture the data lineage necessary to reproduce and investigate predictions. After identifying the cause, the matching correction is: Record governed dataset snapshot and transformation versions as pipeline inputs linked to the resulting model artifact. This matters here because while reproducing a training orchestration incident in isolated staging, the following condition occurs: A model….
Incorrect Answers
A:Inspect component definitions, pipeline package versions, parameters and source revisions used in both environments. This examines pipeline environment parameter promotion rather than the specific mechanism reported here. It is not the decisive first check for model feature lineage for audit. The relevant evidence is: Inspect ML Metadata execution links, table snapshots, transformations and model version lineage.
C:Inspect object prefixes, event filters, completed-batch manifests and recent pipeline execution IDs. This examines eventbridge equivalent event trigger on storage rather than the specific mechanism reported here. It is not the decisive first check for model feature lineage for audit. The relevant evidence is: Inspect ML Metadata execution links, table snapshots, transformations and model version lineage.
D:Inspect incoming schema versions, validation logs, casting rules and downstream prediction anomalies. This examines serving schema validation component rather than the specific mechanism reported here. It is not the decisive first check for model feature lineage for audit. The relevant evidence is: Inspect ML Metadata execution links, table snapshots, transformations and model version lineage.
Question 429
During a production ML inference release on Google Cloud, the following condition occurs: A text embedding model cannot fit into a chosen tiny GPU's memory, but a team only considers increasing container CPU cores. The service must meet its agreed quality and latency requirements. Which targeted engineering change most directly addresses the failure?
- Use stable cohort-level assignment and record model identity on requests and measured outcomes.
- Choose a GPU configuration with sufficient memory or validated model compression while benchmarking cost per query.
- Configure the supported private DNS, network route and endpoint access path without granting public ingress.
- Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses.
Explanation
Correct Answer
B:Choose a GPU configuration with sufficient memory or validated model compression while benchmarking cost per query. This is the targeted corrective response for gpu cost and memory fit for large embedding model. CPU cores do not expand device memory; accelerator selection must match model footprint and request shape. Verify the diagnosis using: Inspect model weights, KV or activation memory, concurrency load, GPU memory, and supported device types. Confirm success using: The model loads at target concurrency without OOM and meets latency and cost targets.
Incorrect Answers
A:Use stable cohort-level assignment and record model identity on requests and measured outcomes. This action addresses a/b assignment consistency for user journeys, not the failure in the described gpu cost and memory fit for large embedding model case. The targeted correction is: Choose a GPU configuration with sufficient memory or validated model compression while benchmarking cost per query.
C:Configure the supported private DNS, network route and endpoint access path without granting public ingress. This action addresses private vpc routing failure for serving, not the failure in the described gpu cost and memory fit for large embedding model case. The targeted correction is: Choose a GPU configuration with sufficient memory or validated model compression while benchmarking cost per query.
D:Apply bounded concurrency, idempotent retries with exponential backoff and explicit overload responses. This action addresses predictive serving rate limit and backpressure, not the failure in the described gpu cost and memory fit for large embedding model case. The targeted correction is: Choose a GPU configuration with sufficient memory or validated model compression while benchmarking cost per query.
Question 430
During model-training validation on a Google Cloud experimental project, 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. The change should preserve a meaningful independent model-quality comparison. Which statement identifies the primary reason for this failure?
- Structured numeric forecasting relies on temporal evidence; a conversational model is not automatically the best forecaster.
- When a model fits on each worker, partitioning training data is often more suitable than splitting the model itself.
- Interpretability is an engineering requirement that must be considered with predictive power, not appended as a label.
- Structured tabular datasets often have competitive simple baselines; architectural complexity is not inherently better.
Explanation
Correct Answer
D:Structured tabular datasets often have competitive simple baselines; architectural complexity is not inherently better. This explains the mechanism behind deep model versus gradient boosted trees on tabular and why the described failure is possible. A discriminating investigation is: Inspect feature modality, example counts, training cost, validation scores and latency requirements. The corresponding repair is: Benchmark a suitable boosted-tree model against more complex candidates and retain complexity only when justified by held-out gains.
Incorrect Answers
A:Structured numeric forecasting relies on temporal evidence; a conversational model is not automatically the best forecaster. This principle describes forecast trend and seasonality versus llm, which is not the cause of the observed deep model versus gradient boosted trees on tabular failure. The controlling explanation is: Structured tabular datasets often have competitive simple baselines; architectural complexity is not inherently better.
B:When a model fits on each worker, partitioning training data is often more suitable than splitting the model itself. This principle describes distributed data parallel gradient synchronization, which is not the cause of the observed deep model versus gradient boosted trees on tabular failure. The controlling explanation is: Structured tabular datasets often have competitive simple baselines; architectural complexity is not inherently better.
C:Interpretability is an engineering requirement that must be considered with predictive power, not appended as a label. This principle describes interpretability-constrained model baseline, which is not the cause of the observed deep model versus gradient boosted trees on tabular failure. The controlling explanation is: Structured tabular datasets often have competitive simple baselines; architectural complexity is not inherently better.