Databricks Certified Machine Learning Professional Practice Test 61
Question 600
The engineering team encounters the following issue: A drift alert fires but its message contains only the word anomaly without dataset, model version or baseline window. What investigation would most effectively isolate this behavior?
- Inspect received alert fields and the corresponding monitor query, model tag and comparison window.
- Review the metric query's output grain, fields and result-table access against data classification.
- Review rolling-window statistics alongside a fixed approved baseline for the same cohort.
- Compare row-level model-version IDs and profile grouping behavior across known rollout periods.
Explanation
Correct Answer
A:This is the most discriminating check for the observed symptoms. Inspect received alert fields and the corresponding monitor query, model tag and comparison window. The underlying technical issue is that alert payloads should carry sufficient metric and lineage context for safe triage.
Incorrect Answers
B:That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A monitor owner writes a custom metric that logs unaggregated account-level health data in a broadly shared results table. For the stated incident, the decisive observation is inspect received alert fields and the corresponding monitor query, model tag and comparison window.
C:That diagnostic evidence 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. For the stated incident, the decisive observation is inspect received alert fields and the corresponding monitor query, model tag and comparison window.
D:That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: An inference profile groups two active model versions into one aggregate and hides a regression introduced by the new release. For the stated incident, the decisive observation is inspect received alert fields and the corresponding monitor query, model tag and comparison window.
Question 601
Consider the following fictional Databricks production case: A tabular classifier fits comfortably on one GPU and 100 million independent examples are available, but a team proposes splitting the small model's layers across hosts. Which result is the strongest acceptance criterion after remediation?
- Re-evaluated candidate trials under the same fixed protocol yield a defensible ranking.
- A controlled A/B benchmark demonstrates the chosen configuration's cost-performance advantage.
- Measured samples per second improve while the model converges consistently with the single-worker baseline.
- A restart retains finished trial metrics and resumes remaining work without rerunning accepted completions.
Explanation
Correct Answer
C:This verification measures the required result rather than just process liveness. Measured samples per second improve while the model converges consistently with the single-worker baseline. It confirms whether the targeted action succeeded: use supported data parallelism with appropriately batched examples and convergence checks.
Incorrect Answers
A:That acceptance test is directed at a different scaling and tuning situation: A hyperparameter team compares two Optuna studies with different train/validation splits and chooses the lower loss as though both metrics were directly comparable. Here the actual success criterion is measured samples per second improve while the model converges consistently with the single-worker baseline.
B:That acceptance test is directed at a different scaling and tuning situation: A moderately sized dataset fits on one worker but a team provisions a huge distributed cluster and observes overhead higher than model fit. Here the actual success criterion is measured samples per second improve while the model converges consistently with the single-worker baseline.
D:That acceptance test is directed at a different scaling and tuning situation: An interrupted Ray hyperparameter sweep has an expensive completed-trial history but a new launch discards its experiment checkpoint. Here the actual success criterion is measured samples per second improve while the model converges consistently with the single-worker baseline.
Question 602
In an enterprise ML workflow, the reported issue is: An automated retraining job triggers whenever a numerical feature mean shifts by 0.1, regardless of normal seasonal changes or sample size. Which change addresses the real cause without undermining other requirements?
- Require a meaningful quality improvement supported by adequate evidence before changing the champion.
- Evaluate all eligible candidates against one controlled frozen benchmark and rank only comparable results.
- Reject candidates that violate required safety segment thresholds even when aggregate F1 increases.
- Gate retraining on validated drift or predictive performance signals with cooldown and approval rules.
Explanation
Correct Answer
D:This corrects the relevant failure boundary instead of merely hiding the symptom. Gate retraining on validated drift or predictive performance signals with cooldown and approval rules. The issue matters because normal weekly seasonality launches wasteful retraining runs and noisy alerts.
Incorrect Answers
A:That corrective action is directed at a different automated retraining situation: A new model beats the champion by 0.1 percentage points on a tiny evaluation set, and the team automatically promotes it despite large uncertainty. The relevant corrective action here is gate retraining on validated drift or predictive performance signals with cooldown and approval rules.
B:That corrective action is directed at a different automated retraining situation: A retraining pipeline ranks candidate models using metrics calculated on different dataset versions, then selects the numerically highest result. The relevant corrective action here is gate retraining on validated drift or predictive performance signals with cooldown and approval rules.
C:That corrective action is directed at a different automated retraining situation: A new fraud classifier improves overall F1 but doubles false negatives for high-value payment accounts, and an automated policy chooses it on F1 alone. The relevant corrective action here is gate retraining on validated drift or predictive performance signals with cooldown and approval rules.
Question 603
A production ML review contains this situation: A bank requires an immediate rollback after deployment and wants the prior serving environment intact until the new environment passes tests. What is the most material immediate risk in this situation?
- High-latency candidate traffic continues despite a registry metadata change.
- The team cannot recover from a bad promotion without an emergency rebuild.
- Older clients receive validation errors even though new client requests succeed.
- Rebuilding the old model under outage pressure extends an avoidable recovery window.
Explanation
Correct Answer
D:This consequence follows from the specific unresolved condition. Rebuilding the old model under outage pressure extends an avoidable recovery window. The causal constraint is: Blue-green deployment keeps two release environments available so traffic can be switched between validated states.
Incorrect Answers
A:That direct consequence is directed at a different deployment strategies situation: During a controlled release, the candidate model's P95 latency triples and engineers want to change the registered model alias without adjusting endpoint routing. The more immediate risk here is rebuilding the old model under outage pressure extends an avoidable recovery window.
B:That direct consequence is directed at a different deployment strategies situation: An ML operator promotes a new model version but its automated rollback points to a model that was deleted from the endpoint configuration. The more immediate risk here is rebuilding the old model under outage pressure extends an avoidable recovery window.
C:That direct consequence is directed at a different deployment strategies situation: A rollout updates the serving model but a renamed input column breaks requests from an older mobile client. The more immediate risk here is rebuilding the old model under outage pressure extends an avoidable recovery window.
Question 604
Review this production model incident: A churn feature job succeeds and a classifier achieves acceptable F1, but the deployed endpoint returns HTTP 400 because request column names differ. What end-to-end verification is most persuasive here?
- The exact saved fixture is successfully scored via the deployed endpoint with expected fields.
- A real staging integration catches schema mismatches that the mocked tests intentionally cannot test.
- The production image digest matches the tested release and reproduces expected fixture outputs.
- Multiple clean sessions produce the same approved encoded indices for every fixture category.
Explanation
Correct Answer
A:This verification measures the required result rather than just process liveness. The exact saved fixture is successfully scored via the deployed endpoint with expected fields. It confirms whether the targeted action succeeded: build an end-to-end integration fixture that verifies schemas, versions and prediction outputs across stages.
Incorrect Answers
B:That acceptance test is directed at a different validation testing situation: A test suite mocks every external interface so aggressively that even an incompatible feature table schema appears valid. Here the actual success criterion is the exact saved fixture is successfully scored via the deployed endpoint with expected fields.
C:That acceptance test 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. Here the actual success criterion is the exact saved fixture is successfully scored via the deployed endpoint with expected fields.
D:That acceptance test is directed at a different validation testing situation: A notebook's label encoder produces a different category order depending on an unordered Python set, but tests assert only output length. Here the actual success criterion is the exact saved fixture is successfully scored via the deployed endpoint with expected fields.
Question 605
Review this production model incident: An inference pipeline sends new categorical labels unseen during training to StringIndexerModel with error handling configured to reject unknown categories. What evidence would best distinguish the actual fault from neighboring ML concerns?
- Inspect fitted category mappings, handleInvalid behavior and a sample newly introduced category.
- Inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values.
- Inspect class prevalence, the confusion matrix, minority recall and false-negative cost rather than headline accuracy.
- Inspect the stage class API and whether fitting produces any data-dependent parameters.
Explanation
Correct Answer
A:This is the most discriminating check for the observed symptoms. Inspect fitted category mappings, handleInvalid behavior and a sample newly introduced category. The underlying technical issue is that training-fitted preprocessors may reject unseen labels; inference handling must be planned and tested explicitly.
Incorrect Answers
B:That diagnostic evidence is directed at a different using spark ml situation: A SparkML grid-search experiment selects the model with the largest error metric even though the evaluator reports RMSE to minimize. For the stated incident, the decisive observation is inspect fitted category mappings, handleInvalid behavior and a sample newly introduced category.
C:That diagnostic evidence 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. For the stated incident, the decisive observation is inspect fitted category mappings, handleInvalid behavior and a sample newly introduced category.
D:That diagnostic evidence 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. For the stated incident, the decisive observation is inspect fitted category mappings, handleInvalid behavior and a sample newly introduced category.
Question 606
The engineering team encounters the following issue: A team hides all preprocessing logic inside one long notebook, making it hard to test independently without starting a Spark cluster. What direct consequence should the owner prioritize if no fix is made?
- Small logic errors require slow integration jobs for every test iteration.
- Sensitive events leak into lower-trust test assets while making debugging harder.
- A model that looks accurate offline cannot be called successfully by production clients.
- Predictions drift between restarts because identical labels occupy different vector positions.
Explanation
Correct Answer
A:This consequence follows from the specific unresolved condition. Small logic errors require slow integration jobs for every test iteration. The causal constraint is: Separating pure functions from orchestration can make unit tests faster while retaining notebook-driven workflows.
Incorrect Answers
B:That direct consequence 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 more immediate risk here is small logic errors require slow integration jobs for every test iteration.
C:That direct consequence is directed at a different validation testing situation: A churn feature job succeeds and a classifier achieves acceptable F1, but the deployed endpoint returns HTTP 400 because request column names differ. The more immediate risk here is small logic errors require slow integration jobs for every test iteration.
D:That direct consequence is directed at a different validation testing situation: A notebook's label encoder produces a different category order depending on an unordered Python set, but tests assert only output length. The more immediate risk here is small logic errors require slow integration jobs for every test iteration.
Question 607
The engineering team encounters the following issue: An ML engineer manually compares two logistic-regression regularization values on a single holdout but has no reusable SparkML parameter-search plan. Which experiment most directly demonstrates a successful fix?
- A small synthetic record passes every pipeline stage and yields a numeric feature vector of expected dimension.
- A controlled example maps known training categories to stable numeric indices using the fitted model.
- Both known and unknown fixture categories yield the documented safe scoring behavior.
- The tracked run records each parameter map, validation metric and final chosen model before untouched test evaluation.
Explanation
Correct Answer
D:This verification measures the required result rather than just process liveness. The tracked run records each parameter map, validation metric and final chosen model before untouched test evaluation. It confirms whether the targeted action succeeded: build an explicit parameter grid with a compatible CrossValidator or TrainValidationSplit and preserve a final untouched test set.
Incorrect Answers
A:That acceptance test is directed at a different using spark ml situation: An IoT classifier passes raw text categories directly into VectorAssembler ahead of StringIndexer and OneHotEncoder, generating incompatible vector inputs. Here the actual success criterion is the tracked run records each parameter map, validation metric and final chosen model before untouched test evaluation.
B:That acceptance test is directed at a different using spark ml situation: A new Spark developer calls transform on an unfitted StringIndexer instance and expects learned category labels to appear automatically. Here the actual success criterion is the tracked run records each parameter map, validation metric and final chosen model before untouched test evaluation.
C:That acceptance test is directed at a different using spark ml situation: An inference pipeline sends new categorical labels unseen during training to StringIndexerModel with error handling configured to reject unknown categories. Here the actual success criterion is the tracked run records each parameter map, validation metric and final chosen model before untouched test evaluation.
Question 608
The engineering team encounters the following issue: A test suite mocks every external interface so aggressively that even an incompatible feature table schema appears valid. What investigation would most effectively isolate this behavior?
- Compare mock schemas with actual feature-table and serving contracts in a controlled environment.
- Compare timestamp values after ingestion, feature-table write and training-set as-of join for one known event.
- Repeat the same synthetic category fixture in separate processes and compare exact mapping outputs.
- Compare individual function tests with an end-to-end fixture crossing feature lookup and inference boundaries.
Explanation
Correct Answer
A:This is the most discriminating check for the observed symptoms. Compare mock schemas with actual feature-table and serving contracts in a controlled environment. The underlying technical issue is that mocks are appropriate for isolated logic but cannot prove real integration contracts.
Incorrect Answers
B:That diagnostic evidence is directed at a different validation testing situation: A feature-refresh task writes rows with event time in local timezone while the training join assumes UTC; unit tests of each component pass. For the stated incident, the decisive observation is compare mock schemas with actual feature-table and serving contracts in a controlled environment.
C:That diagnostic evidence is directed at a different validation testing situation: A notebook's label encoder produces a different category order depending on an unordered Python set, but tests assert only output length. For the stated incident, the decisive observation is compare mock schemas with actual feature-table and serving contracts in a controlled environment.
D:That diagnostic evidence is directed at a different validation testing situation: A team unit-tests each model component independently but its feature table and serving endpoint disagree on column names. For the stated incident, the decisive observation is compare mock schemas with actual feature-table and serving contracts in a controlled environment.
Question 609
In an enterprise ML workflow, the reported issue is: A training dataset combines a feature table with multiple timestamped rows per device using an equality join on device ID only. Which change addresses the real cause without undermining other requirements?
- Use deterministic merge or overwrite semantics scoped to logical feature keys and verify retry behavior.
- Configure time-series columns and a timestamp lookup in the supported feature training-set API.
- Configure an appropriate supported online table and governed feature lookup for the endpoint.
- Use a supported cached or precomputed feature strategy or bounded fallback consistent with freshness needs.
Explanation
Correct Answer
B:This corrects the relevant failure boundary instead of merely hiding the symptom. Configure time-series columns and a timestamp lookup in the supported feature training-set API. The issue matters because training labels duplicate and features from future measurements become attached to older events.
Incorrect Answers
A:That corrective action is directed at a different advanced feature store concepts situation: A feature refresh writes duplicate records for each entity and logical timestamp after retries, breaking unique lookups. The relevant corrective action here is configure time-series columns and a timestamp lookup in the supported feature training-set API.
C:That corrective action is directed at a different advanced feature store concepts situation: A pricing endpoint queries a large offline Delta feature table for every user request, and P95 response time reaches several seconds. The relevant corrective action here is configure time-series columns and a timestamp lookup in the supported feature training-set API.
D:That corrective action is directed at a different advanced feature store concepts situation: A streaming recommender's request-time feature function calls a slow third-party API synchronously on every prediction. The relevant corrective action here is configure time-series columns and a timestamp lookup in the supported feature training-set API.
Question 610
The engineering team encounters the following issue: A training notebook reports F1 for several models but only logs the final model's parameters in MLflow, leaving metrics and hyperparameters disconnected. What investigation would most effectively isolate this behavior?
- Compare original search ID, resumed run tags and candidate-set coverage after restoration.
- Compare run parameter maps against logged evaluation metrics and the local experiment output.
- Inspect logged model environment files and serving import errors for the missing package.
- Compare actual serving request schema, feature lookup capabilities and the model's predict signature.
Explanation
Correct Answer
B:This is the most discriminating check for the observed symptoms. Compare run parameter maps against logged evaluation metrics and the local experiment output. The underlying technical issue is that mLflow run-level logging should attach parameters and evaluated metrics to the corresponding candidate run.
Incorrect Answers
A:That diagnostic evidence is directed at a different advanced mlflow usage situation: A research team restarts a long hyperparameter sweep and merges all newly executed trials into a new experiment without connecting them to the original parent. For the stated incident, the decisive observation is compare run parameter maps against logged evaluation metrics and the local experiment output.
C:That diagnostic evidence is directed at a different advanced mlflow usage situation: A team logs a custom Python model that imports a package installed only in the author's notebook, and the serving endpoint fails when loading the artifact. For the stated incident, the decisive observation is compare run parameter maps against logged evaluation metrics and the local experiment output.
D:That diagnostic evidence 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. For the stated incident, the decisive observation is compare run parameter maps against logged evaluation metrics and the local experiment output.