Databricks Certified Machine Learning Professional Practice Test 80
Question 790
Review this production model incident: A model serving service uses a broadly privileged shared token to read every online feature table when it needs only one approved feature set. What end-to-end verification is most persuasive here?
- The model input is missing or flagged stale when no eligible feature exists inside the authorized window.
- A restart restores feature values consistent with the pre-restart control dataset.
- Authorized prediction requests work while a test access to an unrelated feature table is denied.
- A past labeled event can still retrieve the correct historical feature state.
Explanation
Correct Answer
C:This verification measures the required result rather than just process liveness. Authorized prediction requests work while a test access to an unrelated feature table is denied. It confirms whether the targeted action succeeded: assign a restricted serving identity and permit access only to required approved online feature assets.
Incorrect Answers
A:That acceptance test is directed at a different advanced feature store concepts situation: A sensor classifier silently reuses features from a year-old device after months without readings, despite a maximum allowed lookback of seven days. Here the actual success criterion is authorized prediction requests work while a test access to an unrelated feature table is denied.
B:That acceptance test is directed at a different advanced feature store concepts situation: A business deploys a model whose features are computed from a stream in memory only and disappear when its compute cluster restarts. Here the actual success criterion is authorized prediction requests work while a test access to an unrelated feature table is denied.
D:That acceptance test is directed at a different advanced feature store concepts situation: A model's feature generation step overwrites all feature-table history daily even though time-aware training requires earlier snapshots. Here the actual success criterion is authorized prediction requests work while a test access to an unrelated feature table is denied.
Question 791
Consider the following fictional Databricks production case: A team edits a model endpoint through the UI while its CI pipeline concurrently applies an older configuration and reverts the intended traffic split. What underlying principle determines the most appropriate interpretation?
- Endpoint deployment requires controlled ownership and conflict-safe update/versioning practices.
- Inference clients should honor rate limits and use bounded exponential backoff with idempotent retry semantics.
- MLflow custom-model portability depends on the declared runtime dependencies as well as artifact bytes.
- Serving responses that are cached or audited need model-version provenance and appropriate expiration rules.
Explanation
Correct Answer
A:The technical constraint is specific to the reported failure. Endpoint deployment requires controlled ownership and conflict-safe update/versioning practices. To test it, review endpoint change history, CI deployment commits and the order of conflicting updates.
Incorrect Answers
B:That underlying principle is directed at a different custom model serving situation: A real-time client retries every 429 rate-limit error immediately in an unbounded loop and multiplies load on the serving endpoint. The current incident instead turns on endpoint deployment requires controlled ownership and conflict-safe update/versioning practices.
C:That underlying principle is directed at a different custom model serving situation: A team logs custom MLflow model weights but forgets to include required Python package versions in the model environment. The current incident instead turns on endpoint deployment requires controlled ownership and conflict-safe update/versioning practices.
D:That underlying principle is directed at a different custom model serving situation: An integration caches predictions without recording the model version and returns yesterday's values after a controlled rollback. The current incident instead turns on endpoint deployment requires controlled ownership and conflict-safe update/versioning practices.
Question 792
A production ML review contains this situation: An endpoint's error rate spikes while data profiles look normal, and responders attempt to retrain the model instead of checking serving infrastructure. Which statement best explains the failure mode?
- Alert payloads should carry sufficient metric and lineage context for safe triage.
- Time-series data profiles use a timestamp and granularity to compare measurements across periods.
- Model health, performance and data drift are different failure categories that need separate investigation paths.
- A fixed baseline must be versioned and immutable for the intended comparison period.
Explanation
Correct Answer
C:The technical constraint is specific to the reported failure. Model health, performance and data drift are different failure categories that need separate investigation paths. To test it, inspect endpoint status codes, resource saturation and model-load logs together with current profile metrics.
Incorrect Answers
A:That underlying principle is directed at a different drift detection and lakehouse monitoring situation: A drift alert fires but its message contains only the word anomaly without dataset, model version or baseline window. The current incident instead turns on model health, performance and data drift are different failure categories that need separate investigation paths.
B:That underlying principle is directed at a different drift detection and lakehouse monitoring situation: A fleet sensor table records measurement timestamps and analysts need daily drift trends; a snapshot-only profile erases the time dimension. The current incident instead turns on model health, performance and data drift are different failure categories that need separate investigation paths.
D:That underlying principle is directed at a different drift detection and lakehouse monitoring situation: A baseline table is overwritten every night with the current day's data before a drift calculation, making the comparison always look unchanged. The current incident instead turns on model health, performance and data drift are different failure categories that need separate investigation paths.
Question 793
Consider the following fictional Databricks production case: A media group has a specialized deep neural model that is not implemented as a SparkML estimator and assumes SparkML will automatically distribute its GPU training. Which implementation is most appropriate for the stated incident?
- Use Tokenizer.transform in the correct pipeline position while reserving fit for actual Estimators.
- Select a supported distributed training framework for that model or a deliberate single-node fit rather than relabeling it SparkML.
- Split the raw observations first, then fit every learned preprocessing stage on training data only.
- Evaluate the model with class-sensitive metrics and choose thresholds aligned with false-positive and missed-fraud tradeoffs.
Explanation
Correct Answer
B:This corrects the relevant failure boundary instead of merely hiding the symptom. Select a supported distributed training framework for that model or a deliberate single-node fit rather than relabeling it SparkML. The issue matters because the experiment may run entirely on the driver and underutilize provisioned accelerators.
Incorrect Answers
A:That corrective action is directed at a different using spark ml situation: A text pipeline repeatedly fits a Tokenizer stage even though tokenization simply converts strings into token arrays without learning a model. The relevant corrective action here is select a supported distributed training framework for that model or a deliberate single-node fit rather than relabeling it SparkML.
C:That corrective action is directed at a different using spark ml situation: A credit-risk pipeline fits StringIndexer on the entire dataset before splitting train and holdout, and rare holdout labels influence its learned category mapping. The relevant corrective action here is select a supported distributed training framework for that model or a deliberate single-node fit rather than relabeling it SparkML.
D:That corrective action is directed at a different using spark ml situation: A fraud detector predicts the majority no-fraud class for 99.4% of cases and advertises 99.4% accuracy while missing most positives. The relevant corrective action here is select a supported distributed training framework for that model or a deliberate single-node fit rather than relabeling it SparkML.
Question 794
In an enterprise ML workflow, the reported issue is: A team hides all preprocessing logic inside one long notebook, making it hard to test independently without starting a Spark cluster. How should the engineer validate the correction?
- Known positive and negative synthetic inputs map to the intended serving output labels.
- Most preprocessing tests pass offline without a cluster and the integration notebook still works.
- The defect is caught by development unit tests before any deployment and production smoke remains clean.
- A constant-vector unit fixture returns a documented safe result with no NaN or infinity.
Explanation
Correct Answer
B:This verification measures the required result rather than just process liveness. Most preprocessing tests pass offline without a cluster and the integration notebook still works. It confirms whether the targeted action succeeded: extract testable pure functions into a reusable module and keep notebook orchestration thin.
Incorrect Answers
A:That acceptance test is directed at a different validation testing situation: An end-to-end validation uses only a synthetic positive label and fails to detect a swapped positive/negative class mapping in inference. Here the actual success criterion is most preprocessing tests pass offline without a cluster and the integration notebook still works.
C:That acceptance test is directed at a different validation testing situation: A developer changes a feature transform in a notebook and directly uses production data as the only test case. Here the actual success criterion is most preprocessing tests pass offline without a cluster and the integration notebook still works.
D:That acceptance test is directed at a different validation testing situation: A preprocessing function divides by a standard deviation and returns infinity for a constant numeric column. Here the actual success criterion is most preprocessing tests pass offline without a cluster and the integration notebook still works.
Question 795
A production ML review contains this situation: A SparkML grid-search experiment selects the model with the largest error metric even though the evaluator reports RMSE to minimize. Which diagnostic check provides the strongest direct evidence?
- Check the chosen model library and training API for documented distributed support and GPU worker topology.
- Inspect its declared output column and confirm the downstream estimator consumes a vector rather than predicted labels.
- Inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values.
- Trace when the indexer fit was called relative to the train/validation split and compare observed label inventories.
Explanation
Correct Answer
C:This is the most discriminating check for the observed symptoms. Inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values. The underlying technical issue is that parameter search must use the correct evaluator direction for the chosen loss or utility metric.
Incorrect Answers
A:That diagnostic evidence is directed at a different using spark ml situation: A media group has a specialized deep neural model that is not implemented as a SparkML estimator and assumes SparkML will automatically distribute its GPU training. For the stated incident, the decisive observation is inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values.
B:That diagnostic evidence is directed at a different using spark ml situation: A team wants to convert several numeric inputs into one features vector but treats VectorAssembler as a clustering Estimator that should output predictions. For the stated incident, the decisive observation is inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values.
D:That diagnostic evidence is directed at a different using spark ml situation: A credit-risk pipeline fits StringIndexer on the entire dataset before splitting train and holdout, and rare holdout labels influence its learned category mapping. For the stated incident, the decisive observation is inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values.
Question 796
Consider the following fictional Databricks production case: An endpoint's error rate spikes while data profiles look normal, and responders attempt to retrain the model instead of checking serving infrastructure. Which result is the strongest acceptance criterion after remediation?
- A controlled miscalibrated fixture shows a confidence-bin discrepancy while a calibrated control does not.
- Precision and recall populate only for complete labeled cohorts, while unlabeled drift remains separately reported.
- A synthetic slow drift triggers the fixed-baseline comparison before the model exceeds tolerance.
- An injected 5xx serving fault triggers an availability alert while the drift metric remains independently interpretable.
Explanation
Correct Answer
D:This verification measures the required result rather than just process liveness. An injected 5xx serving fault triggers an availability alert while the drift metric remains independently interpretable. It confirms whether the targeted action succeeded: route availability errors to serving reliability remediation, not to automatic training.
Incorrect Answers
A:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A binary model's average predicted probability remains 0.4, but outcomes show that 0.8 predictions are now correct only half the time. Here the actual success criterion is an injected 5xx serving fault triggers an availability alert while the drift metric remains independently interpretable.
B:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A model's inference table records output probabilities but omits ground-truth labels, yet reviewers expect immediate precision and recall metrics. Here the actual success criterion is an injected 5xx serving fault triggers an availability alert while the drift metric remains independently interpretable.
C:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A streaming payments monitor always compares today with the previous hour, and a slow week-long distribution drift escapes because adjacent hours differ only slightly. Here the actual success criterion is an injected 5xx serving fault triggers an availability alert while the drift metric remains independently interpretable.
Question 797
Consider the following fictional Databricks production case: A feature function maps a null purchase amount to zero without distinguishing missing observations, and a notebook unit test covers only positive amounts. What consequence most specifically follows from leaving the issue unchanged?
- Operators believe a new model is live while the previous model still receives all traffic.
- Production jobs waste resources or mutate test artifacts during business scoring.
- Unknown purchase amounts are silently interpreted as actual zero spending.
- A library compatibility defect escapes despite green tests against an older environment.
Explanation
Correct Answer
C:This consequence follows from the specific unresolved condition. Unknown purchase amounts are silently interpreted as actual zero spending. The causal constraint is: Unit tests should exercise edge cases and expected output contracts, not just the happy-path fixture.
Incorrect Answers
A:That direct consequence is directed at a different validation testing situation: A model version is registered but the deployment job points to an old alias after promotion and continues serving last week's model. The more immediate risk here is unknown purchase amounts are silently interpreted as actual zero spending.
B:That direct consequence is directed at a different validation testing situation: An ML project stores test files directly in each production notebook and execution of a deployed notebook inadvertently reruns expensive tests. The more immediate risk here is unknown purchase amounts are silently interpreted as actual zero spending.
D:That direct consequence is directed at a different validation testing situation: A release pipeline runs development unit tests against a different dependency image from the one deployed to production. The more immediate risk here is unknown purchase amounts are silently interpreted as actual zero spending.
Question 798
In an enterprise ML workflow, the reported issue is: A team records experiments but has no controlled path from a selected registered model to scheduled scoring jobs. What technical distinction is essential to this investigation?
- Environment transitions must preserve or explicitly test library and runtime dependencies as well as source code.
- Tracking, registry, jobs and serving are distinct lifecycle capabilities that must be linked by artifact identity.
- Release environment promotion must separate code portability from data-governance and identity bindings.
- A deploy-code strategy promotes a tested, versioned code artifact through environments rather than editing production independently.
Explanation
Correct Answer
B:The technical constraint is specific to the reported failure. Tracking, registry, jobs and serving are distinct lifecycle capabilities that must be linked by artifact identity. To test it, inventory model run IDs, registered versions, job configuration and deployment references.
Incorrect Answers
A:That underlying principle is directed at a different model lifecycle management situation: A model is promoted to production with its code unchanged but a different unpinned training library version is installed, altering prediction preprocessing. The current incident instead turns on tracking, registry, jobs and serving are distinct lifecycle capabilities that must be linked by artifact identity.
C:That underlying principle is directed at a different model lifecycle management situation: A team trains on curated development features but deploys the same job with credentials that accidentally read unapproved production source tables. The current incident instead turns on tracking, registry, jobs and serving are distinct lifecycle capabilities that must be linked by artifact identity.
D:That underlying principle is directed at a different model lifecycle management situation: A regulated company changes deployed Python notebook code directly in production to correct a model transform while its test environment runs a different untracked revision. The current incident instead turns on tracking, registry, jobs and serving are distinct lifecycle capabilities that must be linked by artifact identity.
Question 799
Review this production model incident: A feature function maps a null purchase amount to zero without distinguishing missing observations, and a notebook unit test covers only positive amounts. Which mitigation resolves the defect rather than concealing it?
- Centralize versioned validation functions and test behavior under all consuming environments.
- Write explicit notebook unit tests for missingness and correct zero handling, then fix the transformation if needed.
- Include paired known-label fixtures and assert prediction class semantics and output ordering.
- Use approved synthetic or appropriately deidentified test data with controlled production smoke verification.
Explanation
Correct Answer
B:This corrects the relevant failure boundary instead of merely hiding the symptom. Write explicit notebook unit tests for missingness and correct zero handling, then fix the transformation if needed. The issue matters because unknown purchase amounts are silently interpreted as actual zero spending.
Incorrect Answers
A:That corrective action is directed at a different validation testing situation: An engineering group duplicates a validation helper across six notebooks, and two copies enforce outdated threshold values. The relevant corrective action here is write explicit notebook unit tests for missingness and correct zero handling, then fix the transformation if needed.
C:That corrective action is directed at a different validation testing situation: An end-to-end validation uses only a synthetic positive label and fails to detect a swapped positive/negative class mapping in inference. The relevant corrective action here is write explicit notebook unit tests for missingness and correct zero handling, then fix the transformation if needed.
D:That corrective action is directed at a different validation testing situation: An engineer uses only production replay data to validate a code change and exposes customer events in an unrestricted test workspace. The relevant corrective action here is write explicit notebook unit tests for missingness and correct zero handling, then fix the transformation if needed.
Question 800
The engineering team encounters the following issue: A training orchestrator starts parallel child fits but all workers accidentally log into one active MLflow run and overwrite each other's metrics. Which experiment most directly demonstrates a successful fix?
- Each worker's fit results appear under its own child run without overwritten artifacts.
- Re-executing the recorded configuration reproduces the tracked validation metric within expected variance.
- A synthetic customer-ID request yields the expected governed feature lookup and versioned prediction.
- A time-bounded replay produces understood feature differences and acceptable model performance.
Explanation
Correct Answer
A:This verification measures the required result rather than just process liveness. Each worker's fit results appear under its own child run without overwritten artifacts. It confirms whether the targeted action succeeded: assign distinct nested run IDs to workers and link them to an explicitly tracked parent experiment.
Incorrect Answers
B:That acceptance test is directed at a different advanced mlflow usage situation: A training notebook reports F1 for several models but only logs the final model's parameters in MLflow, leaving metrics and hyperparameters disconnected. Here the actual success criterion is each worker's fit results appear under its own child run without overwritten artifacts.
C:That acceptance test is directed at a different advanced mlflow usage situation: A real-time churn endpoint accepts customer IDs and must fetch current allowed account features, but a custom model class instead expects the caller to supply the entire internal training row. Here the actual success criterion is each worker's fit results appear under its own child run without overwritten artifacts.
D:That acceptance test is directed at a different advanced mlflow usage situation: A deployed custom model fetches latest feature values during serving while offline evaluation used event-time snapshots without checking the difference. Here the actual success criterion is each worker's fit results appear under its own child run without overwritten artifacts.