Question 270A data science organization loses track of model artifacts when multiple experiments share generic filenames. Which Google Cloud machine-learning design choice best satisfies the stated requirement? Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. Use a scalable Dataflow pipeline for stream processing and write governed outputs to suitable Google Cloud storage. Record immutable model artifacts, preprocessing versions, datasets and execution metadata in Agent Platform ML Metadata. Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata. ExplanationCorrect Answer D:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. To validate the correction: Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. Other Answers A:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. It does not resolve the specific failure; the targeted correction is: Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata. B:Managed distributed stream processing supports windowing, parallelism and large data volumes beyond one Python process. It does not resolve the specific failure; the targeted correction is: Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata. C:Model lineage connects an output model to the data, code and parameters that created it, not just its display name. It does not resolve the specific failure; the targeted correction is: Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata.
Question 271A data science organization loses track of model artifacts when multiple experiments share generic filenames. Which technical mechanism is most relevant to this outcome? Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. Unpinned libraries and runtimes can change numerical behavior or package APIs even when the notebook source is unchanged. Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. Model lineage connects an output model to the data, code and parameters that created it, not just its display name. ExplanationCorrect Answer A:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. This supports the specific implementation: Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata. A suitable independent check is: Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. Other Answers B:Unpinned libraries and runtimes can change numerical behavior or package APIs even when the notebook source is unchanged. The cause in the stated scenario is different: Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. C:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. The cause in the stated scenario is different: Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. D:Model lineage connects an output model to the data, code and parameters that created it, not just its display name. The cause in the stated scenario is different: Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility.
Question 272A data science organization loses track of model artifacts when multiple experiments share generic filenames. Which evidence would best demonstrate that this ML solution meets the requirement? Test throughput, event-time lateness, output correctness and replay behavior against representative streaming fixtures. Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. Inspect confusion matrices and confidence intervals on an untouched holdout with the same evaluation protocol. Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. ExplanationCorrect Answer D:Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. This tests the underlying failure: Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. Other Answers A:Managed distributed stream processing supports windowing, parallelism and large data volumes beyond one Python process. That does not establish the required outcome; verify it by checking: Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. B:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. That does not establish the required outcome; verify it by checking: Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. C:Training accuracy is optimistic and decision thresholds depend on business error costs and representative validation sets. That does not establish the required outcome; verify it by checking: Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage.
Question 273A data science organization loses track of model artifacts when multiple experiments share generic filenames. Which repeatable process control should the ML engineering team adopt? Require every promoted candidate to have a run identifier, source snapshot and recorded evaluation evidence. Enforce source lineage, SQL code review and cost-aware partition predicates in shared analytical pipelines. Retain evaluation prompts, model versions and rubric calibration data to detect regressions after changes. Version pipeline components and test cache invalidation whenever data or model dependencies change. ExplanationCorrect Answer A:Require every promoted candidate to have a run identifier, source snapshot and recorded evaluation evidence. This protects the selected approach because Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. The immediate solution is: Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata. Other Answers B:BigQuery separates managed analytical execution from client memory and supports SQL-native processing of large warehouse datasets. The continuing safeguard required here is: Require every promoted candidate to have a run identifier, source snapshot and recorded evaluation evidence. C:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. The continuing safeguard required here is: Require every promoted candidate to have a run identifier, source snapshot and recorded evaluation evidence. D:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. The continuing safeguard required here is: Require every promoted candidate to have a run identifier, source snapshot and recorded evaluation evidence.
Question 274An agentic assistant performs well on a few demos but its real-world retrieval answer quality is inconsistent. What should the team change to address the specific technical requirement? Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. Pin compatible Python and framework dependencies in a versioned environment with reproducible setup instructions. Log experiments, parameters, metrics and artifact versions in Agent Platform Experiments and ML Metadata. Orchestrate reusable trial stages using Agent Platform Pipelines or Kubeflow Pipelines with caching when safe. ExplanationCorrect Answer A:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. To validate the correction: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. Other Answers B:Unpinned libraries and runtimes can change numerical behavior or package APIs even when the notebook source is unchanged. It does not resolve the specific failure; the targeted correction is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. C:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. It does not resolve the specific failure; the targeted correction is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. D:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. It does not resolve the specific failure; the targeted correction is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate.
Question 275An agentic assistant performs well on a few demos but its real-world retrieval answer quality is inconsistent. Which underlying ML principle makes the proposed approach effective? Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. Managed notebook workspaces make compute and access configurable but do not replace model-code review or dependency governance. Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. ExplanationCorrect Answer D:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. This supports the specific implementation: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. A suitable independent check is: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. Other Answers A:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. The cause in the stated scenario is different: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. B:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. The cause in the stated scenario is different: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. C:Managed notebook workspaces make compute and access configurable but do not replace model-code review or dependency governance. The cause in the stated scenario is different: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance.
Question 276An agentic assistant performs well on a few demos but its real-world retrieval answer quality is inconsistent. What controlled validation most directly confirms the approach worked? Recreate a selected trial using the recorded training snapshot and confirm comparable metrics and artifact lineage. Compare stage execution traces and cached outputs across two runs with unchanged inputs and code. Compare the same feature across online retrieval and a historical cutoff while checking freshness and backfill behavior. Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. ExplanationCorrect Answer D:Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. This tests the underlying failure: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. Other Answers A:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. That does not establish the required outcome; verify it by checking: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. B:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. That does not establish the required outcome; verify it by checking: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. C:Feature Stores can standardize feature definitions across online inference and offline training when point-in-time joins are correct. That does not establish the required outcome; verify it by checking: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples.
Question 277An agentic assistant performs well on a few demos but its real-world retrieval answer quality is inconsistent. Which ongoing standard would make this result reliable for future model versions? Require every promoted candidate to have a run identifier, source snapshot and recorded evaluation evidence. Version pipeline components and test cache invalidation whenever data or model dependencies change. Version stream schemas and establish backpressure, dead-letter and late-data policies for the feature pipeline. Retain evaluation prompts, model versions and rubric calibration data to detect regressions after changes. ExplanationCorrect Answer D:Retain evaluation prompts, model versions and rubric calibration data to detect regressions after changes. This protects the selected approach because Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. The immediate solution is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. Other Answers A:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. The continuing safeguard required here is: Retain evaluation prompts, model versions and rubric calibration data to detect regressions after changes. B:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. The continuing safeguard required here is: Retain evaluation prompts, model versions and rubric calibration data to detect regressions after changes. C:Managed distributed stream processing supports windowing, parallelism and large data volumes beyond one Python process. The continuing safeguard required here is: Retain evaluation prompts, model versions and rubric calibration data to detect regressions after changes.
Question 278A foundation-model evaluator needs systematic groundedness and instruction-following tests before rollout. Which proposed approach best resolves the described model engineering problem? Orchestrate reusable trial stages using Agent Platform Pipelines or Kubeflow Pipelines with caching when safe. Record immutable model artifacts, preprocessing versions, datasets and execution metadata in Agent Platform ML Metadata. Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. Use Gemini Enterprise Agent Platform Feature Store with a governed online serving path and historical feature lineage. ExplanationCorrect Answer C:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. To validate the correction: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. Other Answers A:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. It does not resolve the specific failure; the targeted correction is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. B:Model lineage connects an output model to the data, code and parameters that created it, not just its display name. It does not resolve the specific failure; the targeted correction is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. D:Feature Stores can standardize feature definitions across online inference and offline training when point-in-time joins are correct. It does not resolve the specific failure; the targeted correction is: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate.
Question 279A foundation-model evaluator needs systematic groundedness and instruction-following tests before rollout. Which property of the selected ML workflow accounts for this requirement? Model lineage connects an output model to the data, code and parameters that created it, not just its display name. Local data frames suit modest workloads, while distributed engines add overhead that is justified at larger data scale. Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. ExplanationCorrect Answer C:Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. This supports the specific implementation: Evaluate representative generative AI outputs using task-grounded metrics, human-labeled rubrics and a calibrated LLM judge where appropriate. A suitable independent check is: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. Other Answers A:Model lineage connects an output model to the data, code and parameters that created it, not just its display name. The cause in the stated scenario is different: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. B:Local data frames suit modest workloads, while distributed engines add overhead that is justified at larger data scale. The cause in the stated scenario is different: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. D:Experiment tracking links a training run to settings, evaluation results and produced artifacts for reproducibility. The cause in the stated scenario is different: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance.
Question 280A foundation-model evaluator needs systematic groundedness and instruction-following tests before rollout. Which evidence would best demonstrate that this ML solution meets the requirement? Test throughput, event-time lateness, output correctness and replay behavior against representative streaming fixtures. Compare stage execution traces and cached outputs across two runs with unchanged inputs and code. Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. Inspect confusion matrices and confidence intervals on an untouched holdout with the same evaluation protocol. ExplanationCorrect Answer C:Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. This tests the underlying failure: Generated text requires evaluation beyond ordinary classification accuracy, including factual grounding and task compliance. Other Answers A:Managed distributed stream processing supports windowing, parallelism and large data volumes beyond one Python process. That does not establish the required outcome; verify it by checking: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. B:Pipeline components allow explicit dependencies and reuse of deterministic outputs across training experiments. That does not establish the required outcome; verify it by checking: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples. D:Training accuracy is optimistic and decision thresholds depend on business error costs and representative validation sets. That does not establish the required outcome; verify it by checking: Compare evaluation scores and human adjudication on a frozen multilingual test set with representative failure examples.