Databricks Certified Machine Learning Professional Practice Test 92
Question 910
Consider the following fictional Databricks production case: An analyst loads a fitted LogisticRegressionModel but calls fit on it for every scored batch rather than transform. What consequence most specifically follows from leaving the issue unchanged?
- Complexity and cost increase while the tiny fit becomes slower rather than faster.
- Deployment performance may collapse on customers never seen during optimization.
- Unintended retraining or API failures compromise reproducibility of batch scores.
- A rare new category can fail an entire scoring batch rather than produce a controlled result.
Explanation
Correct Answer
C:This consequence follows from the specific unresolved condition. Unintended retraining or API failures compromise reproducibility of batch scores. The causal constraint is: A fitted Model is a Transformer for scoring; retraining belongs to the corresponding Estimator.
Incorrect Answers
A:That direct consequence is directed at a different using spark ml situation: A laboratory runs a linear regression every quarter on only 280 clean observations and proposes a 48-worker Spark cluster to reduce a two-second local fit. The more immediate risk here is unintended retraining or API failures compromise reproducibility of batch scores.
B:That direct consequence is directed at a different using spark ml situation: A recommender experiment computes all evaluation metrics on training interactions and reports the result as generalization to new customers. The more immediate risk here is unintended retraining or API failures compromise reproducibility of batch scores.
D:That direct consequence 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. The more immediate risk here is unintended retraining or API failures compromise reproducibility of batch scores.
Question 911
The engineering team encounters the following issue: A team updates a shared on-demand feature function without version control and unexpectedly changes both batch retraining and online serving behavior. Which experiment most directly demonstrates a successful fix?
- Two executions of the same input produce identical key counts and values.
- A restart restores feature values consistent with the pre-restart control dataset.
- All registered consumer models pass feature compatibility and expected-output tests against the promoted version.
- Every sampled training feature timestamp is no later than its corresponding prediction cutoff.
Explanation
Correct Answer
C:This verification measures the required result rather than just process liveness. All registered consumer models pass feature compatibility and expected-output tests against the promoted version. It confirms whether the targeted action succeeded: promote a versioned feature definition with backward-compatibility tests and controlled rollout.
Incorrect Answers
A:That acceptance test 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. Here the actual success criterion is all registered consumer models pass feature compatibility and expected-output tests against the promoted version.
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 all registered consumer models pass feature compatibility and expected-output tests against the promoted version.
D:That acceptance test is directed at a different advanced feature store concepts situation: A fraud training row is labeled on March 10 but an ordinary entity join attaches an account balance updated on March 12. Here the actual success criterion is all registered consumer models pass feature compatibility and expected-output tests against the promoted version.
Question 912
In an enterprise ML workflow, the reported issue is: A production alert threshold is hard-coded against last year's baseline, but seasonal traffic has shifted and the owner never reviewed the threshold. How should the engineer validate the correction?
- Two fixtures with equal risk proportions but different volumes produce the same custom rate.
- A synthetic v8-only error increase appears in its own model-version metric.
- A synthetic poor new-version cohort is distinguishable from unchanged champion predictions.
- A known benign seasonal fixture stays below the revised threshold and a material anomaly exceeds it.
Explanation
Correct Answer
D:This verification measures the required result rather than just process liveness. A known benign seasonal fixture stays below the revised threshold and a material anomaly exceeds it. It confirms whether the targeted action succeeded: recalibrate alert limits through reviewed baseline management while retaining real anomaly sensitivity.
Incorrect Answers
A:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A monitoring team wants a business metric for the share of transactions exceeding a risk threshold but records only the raw count of high-risk transactions. Here the actual success criterion is a known benign seasonal fixture stays below the revised threshold and a material anomaly exceeds it.
B:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A scoring endpoint quietly moves from model v7 to v8 during a rollout but an inference trend aggregates both versions in one error curve. Here the actual success criterion is a known benign seasonal fixture stays below the revised threshold and a material anomaly exceeds it.
C:That acceptance test 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. Here the actual success criterion is a known benign seasonal fixture stays below the revised threshold and a material anomaly exceeds it.
Question 913
In an enterprise ML workflow, the reported issue is: A text pipeline repeatedly fits a Tokenizer stage even though tokenization simply converts strings into token arrays without learning a model. What likely operational or model-quality harm does this issue create?
- The workflow fails or includes unnecessary invalid fit calls that cannot create a tokenizer model.
- Pipeline fitting fails with missing-column or mismatched-schema errors.
- Deployment performance may collapse on customers never seen during optimization.
- The same raw category maps to a different vector position and produces misleading predictions.
Explanation
Correct Answer
A:This consequence follows from the specific unresolved condition. The workflow fails or includes unnecessary invalid fit calls that cannot create a tokenizer model. The causal constraint is: Tokenizer is a Transformer and applies transform directly without a separate learned fit step.
Incorrect Answers
B:That direct consequence is directed at a different using spark ml situation: A pipeline engineer fits a vectorizer after an estimator in the stage list and then tries to use the estimator's prediction as the vectorizer input. The more immediate risk here is the workflow fails or includes unnecessary invalid fit calls that cannot create a tokenizer model.
C:That direct consequence is directed at a different using spark ml situation: A recommender experiment computes all evaluation metrics on training interactions and reports the result as generalization to new customers. The more immediate risk here is the workflow fails or includes unnecessary invalid fit calls that cannot create a tokenizer model.
D:That direct consequence is directed at a different using spark ml situation: A batch scoring service loads a classifier version from last month while using today's new feature indexer, silently changing categorical positions. The more immediate risk here is the workflow fails or includes unnecessary invalid fit calls that cannot create a tokenizer model.
Question 914
A production ML review contains this situation: A Databricks SDK deployment script sends a model-version update but does not wait for the endpoint to reach a usable ready state before announcing success. What is the most material immediate risk in this situation?
- Automation falsely reports a completed rollout while clients still receive temporary failures.
- Stale candidate predictions persist after traffic has returned to the champion.
- A malformed request crashes the entire serving worker or produces an untraceable score.
- Clients throw parsing errors or accidentally interpret confidence as the prediction.
Explanation
Correct Answer
A:This consequence follows from the specific unresolved condition. Automation falsely reports a completed rollout while clients still receive temporary failures. The causal constraint is: Endpoint create/update calls and operational readiness are separate lifecycle checkpoints.
Incorrect Answers
B:That direct consequence 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 more immediate risk here is automation falsely reports a completed rollout while clients still receive temporary failures.
C:That direct consequence is directed at a different custom model serving situation: A custom PyFunc's predict method assumes every request contains a numeric feature that is optional under the documented inference schema. The more immediate risk here is automation falsely reports a completed rollout while clients still receive temporary failures.
D:That direct consequence is directed at a different custom model serving situation: A serving client calls the REST endpoint expecting one number but the custom PyFunc returns a structured list of prediction and confidence fields. The more immediate risk here is automation falsely reports a completed rollout while clients still receive temporary failures.
Question 915
A production ML review contains this situation: A travel site serves 30,000 bookings per minute and wants to evaluate a new fraud model while limiting customer impact during its first hour. Which check shows that this particular problem has been resolved?
- An endpoint configuration readback and sampled test traffic confirm the intended distribution.
- A readiness probe and known inference fixture both succeed before planned traffic movement.
- A release rehearsal restores champion predictions using the saved model/version and configuration.
- A simulated candidate failure affects only the canary share and triggers a stop or rollback gate.
Explanation
Correct Answer
D:This verification measures the required result rather than just process liveness. A simulated candidate failure affects only the canary share and triggers a stop or rollback gate. It confirms whether the targeted action succeeded: send a small approved percentage of eligible live requests to the candidate and expand only after defined gates pass.
Incorrect Answers
A:That acceptance test is directed at a different deployment strategies situation: A Databricks Model Serving endpoint hosts champion and candidate versions, but a release config accidentally routes 100% of traffic to the candidate instead of 10%. Here the actual success criterion is a simulated candidate failure affects only the canary share and triggers a stop or rollback gate.
B:That acceptance test is directed at a different deployment strategies situation: An endpoint update request succeeds but the newly configured model version is still loading; monitoring marks the release complete immediately. Here the actual success criterion is a simulated candidate failure affects only the canary share and triggers a stop or rollback gate.
C:That acceptance test 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. Here the actual success criterion is a simulated candidate failure affects only the canary share and triggers a stop or rollback gate.
Question 916
Review this production model incident: During a controlled release, the candidate model's P95 latency triples and engineers want to change the registered model alias without adjusting endpoint routing. What is the central engineering constraint behind this behavior?
- Rollout safety needs operational latency and memory gates alongside predictive quality metrics.
- Rare-outcome evaluation requires sufficient exposure, observation time and mature labels before concluding parity.
- Rollout recovery must address the actual endpoint traffic configuration, not assume an alias automatically changes every active route.
- A canary rollout can expose a small controlled share of live traffic while retaining a proven model for most requests.
Explanation
Correct Answer
C:The technical constraint is specific to the reported failure. Rollout recovery must address the actual endpoint traffic configuration, not assume an alias automatically changes every active route. To test it, compare effective serving traffic split, model identities and alias resolution in the live endpoint.
Incorrect Answers
A:That underlying principle is directed at a different deployment strategies situation: A new recommendation model consumes three times as much memory and the team plans a canary based solely on prediction accuracy, ignoring serving capacity. The current incident instead turns on rollout recovery must address the actual endpoint traffic configuration, not assume an alias automatically changes every active route.
B:That underlying principle is directed at a different deployment strategies situation: A candidate fraud model has near-identical average error to the champion, but an organization wants to compare rare fraudulent transactions during a brief canary. The current incident instead turns on rollout recovery must address the actual endpoint traffic configuration, not assume an alias automatically changes every active route.
D:That underlying principle is directed at a different deployment strategies situation: A travel site serves 30,000 bookings per minute and wants to evaluate a new fraud model while limiting customer impact during its first hour. The current incident instead turns on rollout recovery must address the actual endpoint traffic configuration, not assume an alias automatically changes every active route.
Question 917
Consider the following fictional Databricks production case: A team records experiments but has no controlled path from a selected registered model to scheduled scoring jobs. Which implementation is most appropriate for the stated incident?
- Parameterize approved environment resources and use narrowly scoped credentials per stage.
- Pin and validate supported dependency versions as part of the promoted model/code release.
- Build and validate one immutable release artifact before promoting it through controlled environments.
- Bind training outputs to registered immutable versions and promote them through governed jobs or serving endpoints.
Explanation
Correct Answer
D:This corrects the relevant failure boundary instead of merely hiding the symptom. Bind training outputs to registered immutable versions and promote them through governed jobs or serving endpoints. The issue matters because the wrong experiment artifact may be used in production because lifecycle stages are disconnected.
Incorrect Answers
A:That corrective action 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 relevant corrective action here is bind training outputs to registered immutable versions and promote them through governed jobs or serving endpoints.
B:That corrective action 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 relevant corrective action here is bind training outputs to registered immutable versions and promote them through governed jobs or serving endpoints.
C:That corrective action 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 relevant corrective action here is bind training outputs to registered immutable versions and promote them through governed jobs or serving endpoints.
Question 918
In an enterprise ML workflow, the reported issue is: A single-node model training run spends almost all time swapping a 40 GB feature matrix on a 16 GB VM; no distributed version exists. What technical distinction is essential to this investigation?
- Tuning runs must use equivalent evaluation data and metric definitions before ranking configurations.
- For a single-node algorithm, adequate local RAM or a compatible distributed alternative matters more than adding idle executors.
- Ray task orchestration overhead can outweigh useful work for extremely short optimization trials.
- Spark SQL is often the more direct choice for distributed tabular transformations, while Ray targets flexible task/actor workloads.
Explanation
Correct Answer
B:The technical constraint is specific to the reported failure. For a single-node algorithm, adequate local RAM or a compatible distributed alternative matters more than adding idle executors. To test it, inspect resident memory, paging metrics and the training library's process placement.
Incorrect Answers
A:That underlying principle 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. The current incident instead turns on for a single-node algorithm, adequate local RAM or a compatible distributed alternative matters more than adding idle executors.
C:That underlying principle is directed at a different scaling and tuning situation: A researcher distributes 500 tiny one-second trials across many Ray workers and sees coordination cost exceed compute time. The current incident instead turns on for a single-node algorithm, adequate local RAM or a compatible distributed alternative matters more than adding idle executors.
D:That underlying principle is directed at a different scaling and tuning situation: A team already has a vectorized Spark DataFrame pipeline and proposes converting everything to Ray datasets solely to perform a simple SQL aggregation. The current incident instead turns on for a single-node algorithm, adequate local RAM or a compatible distributed alternative matters more than adding idle executors.
Question 919
Consider the following fictional Databricks production case: A time-series monitor uses a string-formatted date that cannot be reliably converted to timestamps and silently groups most rows into an unknown bucket. Which observation most directly tests the suspected failure boundary?
- Inspect timestamp type, event cadence and intended daily windows before choosing the monitor.
- Test timestamp conversion and timezone interpretation for valid and malformed synthetic source rows.
- Stratify historical samples by day type and examine whether the observed difference is expected seasonality.
- Compare spike duration and alert time window against the approved ten-minute persistence rule.
Explanation
Correct Answer
B:This is the most discriminating check for the observed symptoms. Test timestamp conversion and timezone interpretation for valid and malformed synthetic source rows. The underlying technical issue is that time-series profile configuration needs a valid temporal column with reliable timestamp semantics.
Incorrect Answers
A:That diagnostic evidence 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. For the stated incident, the decisive observation is test timestamp conversion and timezone interpretation for valid and malformed synthetic source rows.
C:That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: An online feature monitor compares weekend shopping to weekday reference data and treats expected weekend behavior as a model failure. For the stated incident, the decisive observation is test timestamp conversion and timezone interpretation for valid and malformed synthetic source rows.
D:That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A monitor alerts whenever a ratio exceeds 0.3 for a single minute, although normal traffic produces brief spikes and operations require ten-minute persistence. For the stated incident, the decisive observation is test timestamp conversion and timezone interpretation for valid and malformed synthetic source rows.
Question 920
Review this production model incident: A research team captures only daily batch inference totals with no row-level timestamps but insists on per-minute live inference drift analysis. Which mitigation resolves the defect rather than concealing it?
- Recalibrate alert limits through reviewed baseline management while retaining real anomaly sensitivity.
- Improve authorized logging and table design before configuring fine-grained inference time-series analysis.
- Separate versioned baseline storage from current data and block unintended updates.
- Route availability errors to serving reliability remediation, not to automatic training.
Explanation
Correct Answer
B:This corrects the relevant failure boundary instead of merely hiding the symptom. Improve authorized logging and table design before configuring fine-grained inference time-series analysis. The issue matters because the monitor fabricates apparent minute-level precision from daily totals and misleads responders.
Incorrect Answers
A:That corrective action is directed at a different drift detection and lakehouse monitoring situation: A production alert threshold is hard-coded against last year's baseline, but seasonal traffic has shifted and the owner never reviewed the threshold. The relevant corrective action here is improve authorized logging and table design before configuring fine-grained inference time-series analysis.
C:That corrective action 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 relevant corrective action here is improve authorized logging and table design before configuring fine-grained inference time-series analysis.
D:That corrective action is directed at a different drift detection and lakehouse monitoring situation: An endpoint's error rate spikes while data profiles look normal, and responders attempt to retrain the model instead of checking serving infrastructure. The relevant corrective action here is improve authorized logging and table design before configuring fine-grained inference time-series analysis.