Databricks Certified Machine Learning Professional Practice Test 63
Question 620
Consider the following fictional Databricks production case: 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. What consequence most specifically follows from leaving the issue unchanged?
- Rare but severe forecast misses can trigger inventory shortages while average reporting looks healthy.
- Complexity and cost increase while the tiny fit becomes slower rather than faster.
- A predictor is given nonexistent features or operators expect labels from a nonpredictive stage.
- The supposedly unseen holdout can affect model inputs and inflate reported generalization.
Explanation
Correct Answer
B:This consequence follows from the specific unresolved condition. Complexity and cost increase while the tiny fit becomes slower rather than faster. The causal constraint is: Small workloads may run more efficiently on one node because Spark startup, scheduling and serialization have fixed costs.
Incorrect Answers
A:That direct consequence is directed at a different using spark ml situation: A forecasting model has a few catastrophic numeric errors hidden by a low median absolute error and operations care about large misses. The more immediate risk here is complexity and cost increase while the tiny fit becomes slower rather than faster.
C:That direct consequence 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. The more immediate risk here is complexity and cost increase while the tiny fit becomes slower rather than faster.
D:That direct consequence 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 more immediate risk here is complexity and cost increase while the tiny fit becomes slower rather than faster.
Question 621
Review this production model incident: A binary model's average predicted probability remains 0.4, but outcomes show that 0.8 predictions are now correct only half the time. What end-to-end verification is most persuasive here?
- A controlled miscalibrated fixture shows a confidence-bin discrepancy while a calibrated control does not.
- A production-like model load stays within memory headroom through full request concurrency.
- A test alert identifies the precise asset, time window, threshold and metric that caused it.
- Matching-filter fixtures show no artificial drift while a real within-segment change remains detectable.
Explanation
Correct Answer
A:This verification measures the required result rather than just process liveness. A controlled miscalibrated fixture shows a confidence-bin discrepancy while a calibrated control does not. It confirms whether the targeted action succeeded: track calibrated outcome frequencies and trigger review when confidence-to-outcome consistency changes.
Incorrect Answers
B:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A new model version uses twice the memory per replica, but capacity monitoring tracks only CPU and crash counts. Here the actual success criterion is a controlled miscalibrated fixture shows a confidence-bin discrepancy while a calibrated control does not.
C:That acceptance test 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. Here the actual success criterion is a controlled miscalibrated fixture shows a confidence-bin discrepancy while a calibrated control does not.
D:That acceptance test is directed at a different drift detection and lakehouse monitoring situation: A customer-risk monitoring job compares an unfiltered current high-risk segment to an all-customer historical baseline and interprets the difference as temporal drift. Here the actual success criterion is a controlled miscalibrated fixture shows a confidence-bin discrepancy while a calibrated control does not.
Question 622
Consider the following fictional Databricks production case: An engineer logs a customer-specific API token inside the pickled PythonModel object before registering it to Unity Catalog. Which result is the strongest acceptance criterion after remediation?
- An artifact scan contains no credential strings while authorized serving requests still work.
- The scoped automation can deploy the authorized model and cannot access unrelated catalog models.
- A staging rebuild from retained configuration reproduces endpoint behavior and access rules.
- A saved two-record request returns correctly ordered predictions with the expected schema.
Explanation
Correct Answer
A:This verification measures the required result rather than just process liveness. An artifact scan contains no credential strings while authorized serving requests still work. It confirms whether the targeted action succeeded: remove embedded secrets and configure an approved scoped secret retrieval mechanism for the serving process.
Incorrect Answers
B:That acceptance test is directed at a different custom model serving situation: A test engineer can create a model-serving endpoint in a sandbox but the production deployment fails because the automation identity lacks registered-model access. Here the actual success criterion is an artifact scan contains no credential strings while authorized serving requests still work.
C:That acceptance test is directed at a different custom model serving situation: A platform engineer creates a custom serving endpoint manually and assumes the exact settings can be recovered after an accidental deletion. Here the actual success criterion is an artifact scan contains no credential strings while authorized serving requests still work.
D:That acceptance test is directed at a different custom model serving situation: A notebook uses the MLflow Deployments SDK to send a column-major payload to a custom model endpoint expecting a different documented dataframe schema. Here the actual success criterion is an artifact scan contains no credential strings while authorized serving requests still work.
Question 623
Consider the following fictional Databricks production case: A batch scoring pipeline trains one per-region model and then joins model predictions back using region alone, duplicating rows because each region contains multiple versioned models. What underlying principle determines the most appropriate interpretation?
- Fault-tolerant distributed tuning should use supported experiment storage and restoration rather than recreating every trial.
- Per-group model predictions require unambiguous region-and-model-version associations before row-level output joins.
- Reliable tuning provenance requires parameter, metric and trial identity logging for every evaluated trial.
- Synchronized distributed training requires a coherent parameter-update protocol and checkpoint versioning.
Explanation
Correct Answer
B:The technical constraint is specific to the reported failure. Per-group model predictions require unambiguous region-and-model-version associations before row-level output joins. To test it, inspect model registry versions, join keys and output cardinality against source record IDs.
Incorrect Answers
A:That underlying principle 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. The current incident instead turns on per-group model predictions require unambiguous region-and-model-version associations before row-level output joins.
C:That underlying principle is directed at a different scaling and tuning situation: An Optuna search runs 80 hyperparameter trials but records only the last winning score in MLflow, making comparisons impossible. The current incident instead turns on per-group model predictions require unambiguous region-and-model-version associations before row-level output joins.
D:That underlying principle is directed at a different scaling and tuning situation: A training team averages gradients from multiple workers but lets each worker use a different untracked version of the model. The current incident instead turns on per-group model predictions require unambiguous region-and-model-version associations before row-level output joins.
Question 624
The engineering team encounters the following issue: A fraud training row is labeled on March 10 but an ordinary entity join attaches an account balance updated on March 12. What direct consequence should the owner prioritize if no fix is made?
- Restarted jobs double count behavior or lose events, corrupting online feature values.
- Credential compromise exposes unrelated customers' private feature data.
- Future account information leaks into historical training labels and inflates validation.
- Training consumes stale or differently computed features depending on who ran the notebook.
Explanation
Correct Answer
C:This consequence follows from the specific unresolved condition. Future account information leaks into historical training labels and inflates validation. The causal constraint is: Point-in-time training joins must select features available at or before each label observation, not the latest record irrespective of date.
Incorrect Answers
A:That direct consequence is directed at a different advanced feature store concepts situation: An online behavioral feature counter consumes a streaming source but does not retain checkpoint state after worker restarts. The more immediate risk here is future account information leaks into historical training labels and inflates validation.
B:That direct consequence is directed at a different advanced feature store concepts situation: A model serving service uses a broadly privileged shared token to read every online feature table when it needs only one approved feature set. The more immediate risk here is future account information leaks into historical training labels and inflates validation.
D:That direct consequence is directed at a different advanced feature store concepts situation: A team manually reruns a feature notebook each week before training; feature definitions change without an execution history or stable destination. The more immediate risk here is future account information leaks into historical training labels and inflates validation.
Question 625
Consider the following fictional Databricks production case: A real-time client retries every 429 rate-limit error immediately in an unbounded loop and multiplies load on the serving endpoint. Which observation most directly tests the suspected failure boundary?
- Compare sandbox/prod identities, model grants and sanitized deployment authorization errors.
- Compare training import versions and logged model environment metadata with serving import failures.
- Inspect the model's artifact map and the load_context path resolution under an isolated inference environment.
- Inspect rate-limit response headers, client retry timeline and duplicate request identifiers.
Explanation
Correct Answer
D:This is the most discriminating check for the observed symptoms. Inspect rate-limit response headers, client retry timeline and duplicate request identifiers. The underlying technical issue is that inference clients should honor rate limits and use bounded exponential backoff with idempotent retry semantics.
Incorrect Answers
A:That diagnostic evidence is directed at a different custom model serving situation: A test engineer can create a model-serving endpoint in a sandbox but the production deployment fails because the automation identity lacks registered-model access. For the stated incident, the decisive observation is inspect rate-limit response headers, client retry timeline and duplicate request identifiers.
B:That diagnostic evidence 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. For the stated incident, the decisive observation is inspect rate-limit response headers, client retry timeline and duplicate request identifiers.
C:That diagnostic evidence is directed at a different custom model serving situation: A custom PyFunc estimator loads a tokenizer file from an engineer's laptop path when called in a serving container. For the stated incident, the decisive observation is inspect rate-limit response headers, client retry timeline and duplicate request identifiers.
Question 626
Review this production model incident: A team hides all preprocessing logic inside one long notebook, making it hard to test independently without starting a Spark cluster. What is the central engineering constraint behind this behavior?
- Unit tests should exercise edge cases and expected output contracts, not just the happy-path fixture.
- Separating pure functions from orchestration can make unit tests faster while retaining notebook-driven workflows.
- Cross-component tests must validate shared time semantics and feature availability throughout the ML pipeline.
- Unit tests belong on controlled synthetic inputs before integrations, and production smoke tests do not replace them.
Explanation
Correct Answer
B:The technical constraint is specific to the reported failure. Separating pure functions from orchestration can make unit tests faster while retaining notebook-driven workflows. To test it, identify deterministic data transformations entangled with session setup and filesystem side effects.
Incorrect Answers
A:That underlying principle is directed at a different validation testing situation: A feature function maps a null purchase amount to zero without distinguishing missing observations, and a notebook unit test covers only positive amounts. The current incident instead turns on separating pure functions from orchestration can make unit tests faster while retaining notebook-driven workflows.
C:That underlying principle 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. The current incident instead turns on separating pure functions from orchestration can make unit tests faster while retaining notebook-driven workflows.
D:That underlying principle 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. The current incident instead turns on separating pure functions from orchestration can make unit tests faster while retaining notebook-driven workflows.
Question 627
The engineering team encounters the following issue: 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. What corrective action should the engineer prefer?
- Gate retraining on validated drift or predictive performance signals with cooldown and approval rules.
- Run candidate training after the trigger but require explicit quality/security gates before any live promotion.
- Restore missing data and validate collection health before deciding whether a model retrain is warranted.
- Reject candidates that violate required safety segment thresholds even when aggregate F1 increases.
Explanation
Correct Answer
D:This corrects the relevant failure boundary instead of merely hiding the symptom. Reject candidates that violate required safety segment thresholds even when aggregate F1 increases. The issue matters because a statistically better average model causes larger losses for the protected operational segment.
Incorrect Answers
A:That corrective action is directed at a different automated retraining situation: An automated retraining job triggers whenever a numerical feature mean shifts by 0.1, regardless of normal seasonal changes or sample size. The relevant corrective action here is reject candidates that violate required safety segment thresholds even when aggregate F1 increases.
B:That corrective action is directed at a different automated retraining situation: An alert from a feature monitor automatically replaces the live model with a newly trained artifact without testing or approval. The relevant corrective action here is reject candidates that violate required safety segment thresholds even when aggregate F1 increases.
C:That corrective action is directed at a different automated retraining situation: A feature stream loses half its source events and a retraining policy treats the resulting input shift as proof that the learned concept changed. The relevant corrective action here is reject candidates that violate required safety segment thresholds even when aggregate F1 increases.
Question 628
A production ML review contains this situation: A researcher distributes 500 tiny one-second trials across many Ray workers and sees coordination cost exceed compute time. Which diagnostic check provides the strongest direct evidence?
- Examine Optuna storage connection logs, trial state transitions and worker collision frequency.
- Measure trial startup, scheduling, serialization and actual objective runtime separately.
- Compare operation requirements, data conversion cost and supported query optimizations in both engines.
- Trace parameter ownership and update communication in the proposed training algorithm.
Explanation
Correct Answer
B:This is the most discriminating check for the observed symptoms. Measure trial startup, scheduling, serialization and actual objective runtime separately. The underlying technical issue is that ray task orchestration overhead can outweigh useful work for extremely short optimization trials.
Incorrect Answers
A:That diagnostic evidence is directed at a different scaling and tuning situation: A researcher parallelizes Optuna trials while sharing one writable SQLite file on a fragile network mount and observes duplicate reservations and storage locks. For the stated incident, the decisive observation is measure trial startup, scheduling, serialization and actual objective runtime separately.
C:That diagnostic evidence 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. For the stated incident, the decisive observation is measure trial startup, scheduling, serialization and actual objective runtime separately.
D:That diagnostic evidence is directed at a different scaling and tuning situation: An ML engineer tries to parallelize a stateful neural-training loop by executing independent mapPartitions calls that share one mutable global model in the driver. For the stated incident, the decisive observation is measure trial startup, scheduling, serialization and actual objective runtime separately.
Question 629
Consider the following fictional Databricks production case: A feature-refresh task writes rows with event time in local timezone while the training join assumes UTC; unit tests of each component pass. Which result is the strongest acceptance criterion after remediation?
- The same event time is interpreted consistently at every persisted lifecycle boundary.
- The integration fixture traverses every component with matching validated field names.
- Multiple clean sessions produce the same approved encoded indices for every fixture category.
- A real staging integration catches schema mismatches that the mocked tests intentionally cannot test.
Explanation
Correct Answer
A:This verification measures the required result rather than just process liveness. The same event time is interpreted consistently at every persisted lifecycle boundary. It confirms whether the targeted action succeeded: add a full-path integration test with explicit time-zone and cutoff assertions.
Incorrect Answers
B:That acceptance test 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. Here the actual success criterion is the same event time is interpreted consistently at every persisted lifecycle boundary.
C: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 same event time is interpreted consistently at every persisted lifecycle boundary.
D: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 same event time is interpreted consistently at every persisted lifecycle boundary.
Question 630
A production ML review contains this situation: A team trains on curated development features but deploys the same job with credentials that accidentally read unapproved production source tables. Which statement best explains the failure mode?
- Experiment tracking metadata is not a substitute for governed asset privileges and serving access control.
- Governed model registration and access controls provide traceability beyond an ad hoc personal file path.
- Release environment promotion must separate code portability from data-governance and identity bindings.
- Tracking, registry, jobs and serving are distinct lifecycle capabilities that must be linked by artifact identity.
Explanation
Correct Answer
C:The technical constraint is specific to the reported failure. Release environment promotion must separate code portability from data-governance and identity bindings. To test it, inspect environment-specific data-source references and effective service-identity permissions.
Incorrect Answers
A:That underlying principle is directed at a different model lifecycle management situation: A platform team proposes using MLflow Tracking alone to enforce production endpoint authorization, ignoring Unity Catalog and service permissions. The current incident instead turns on release environment promotion must separate code portability from data-governance and identity bindings.
B:That underlying principle is directed at a different model lifecycle management situation: An engineer stores model weights in a personal workspace folder and treats that path as a production artifact registry with audit history. The current incident instead turns on release environment promotion must separate code portability from data-governance and identity bindings.
D:That underlying principle is directed at a different model lifecycle management situation: A team records experiments but has no controlled path from a selected registered model to scheduled scoring jobs. The current incident instead turns on release environment promotion must separate code portability from data-governance and identity bindings.