Databricks Certified Machine Learning Professional Practice Test 59
Question 580
Review this production model incident: 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%. Which mitigation resolves the defect rather than concealing it?
- Send a small approved percentage of eligible live requests to the candidate and expand only after defined gates pass.
- Set a statistically and operationally adequate rollout observation period with strict safety abort thresholds.
- Set and verify the intended approved candidate-versus-champion traffic split before exposure.
- Prepare a complete tested green environment and switch traffic after approval while retaining blue for rollback.
Explanation
Correct Answer
C:This corrects the relevant failure boundary instead of merely hiding the symptom. Set and verify the intended approved candidate-versus-champion traffic split before exposure. The issue matters because all requests are exposed to the candidate during what operators believed was a small canary.
Incorrect Answers
A:That corrective action 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 relevant corrective action here is set and verify the intended approved candidate-versus-champion traffic split before exposure.
B:That corrective action 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 relevant corrective action here is set and verify the intended approved candidate-versus-champion traffic split before exposure.
D:That corrective action is directed at a different deployment strategies situation: A bank requires an immediate rollback after deployment and wants the prior serving environment intact until the new environment passes tests. The relevant corrective action here is set and verify the intended approved candidate-versus-champion traffic split before exposure.
Question 581
Consider the following fictional Databricks production case: A test suite mocks every external interface so aggressively that even an incompatible feature table schema appears valid. Which result is the strongest acceptance criterion after remediation?
- A constant-vector unit fixture returns a documented safe result with no NaN or infinity.
- The same event time is interpreted consistently at every persisted lifecycle boundary.
- 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.
Explanation
Correct Answer
C:This verification measures the required result rather than just process liveness. A real staging integration catches schema mismatches that the mocked tests intentionally cannot test. It confirms whether the targeted action succeeded: retain fast mocked unit tests and add realistic integration contract tests for critical boundaries.
Incorrect Answers
A: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 a real staging integration catches schema mismatches that the mocked tests intentionally cannot test.
B:That acceptance test 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. Here the actual success criterion is a real staging integration catches schema mismatches that the mocked tests intentionally cannot test.
D: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 a real staging integration catches schema mismatches that the mocked tests intentionally cannot test.
Question 582
In an enterprise ML workflow, the reported issue is: A notebook's label encoder produces a different category order depending on an unordered Python set, but tests assert only output length. How should the engineer validate the correction?
- Multiple clean sessions produce the same approved encoded indices for every fixture category.
- Test runs require no unauthorized identifiers and negative access probes fail as intended.
- 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.
Explanation
Correct Answer
A:This verification measures the required result rather than just process liveness. Multiple clean sessions produce the same approved encoded indices for every fixture category. It confirms whether the targeted action succeeded: use deterministic mapping order and assert category-to-index values in notebook tests.
Incorrect Answers
B:That acceptance test 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. Here the actual success criterion is multiple clean sessions produce the same approved encoded indices for every fixture category.
C:That acceptance test 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. Here the actual success criterion is multiple clean sessions produce the same approved encoded indices for every fixture category.
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 multiple clean sessions produce the same approved encoded indices for every fixture category.
Question 583
The engineering team encounters the following issue: A pricing endpoint queries a large offline Delta feature table for every user request, and P95 response time reaches several seconds. Which technical principle most directly explains the root of the issue?
- Historical feature timelines must be retained when reproducible point-in-time joins depend on prior values.
- Streaming derived features should deduplicate at a stable business-event key before non-idempotent aggregation.
- Point-in-time training joins must select features available at or before each label observation, not the latest record irrespective of date.
- Low-latency serving requires online feature access with a documented freshness and lookup contract.
Explanation
Correct Answer
D:The technical constraint is specific to the reported failure. Low-latency serving requires online feature access with a documented freshness and lookup contract. To test it, measure endpoint latency attributable to offline table reads versus approved online lookup latency.
Incorrect Answers
A:That underlying principle 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. The current incident instead turns on low-latency serving requires online feature access with a documented freshness and lookup contract.
B:That underlying principle is directed at a different advanced feature store concepts situation: A real-time feature job updates user purchase totals without checking whether upstream transaction events are delivered twice. The current incident instead turns on low-latency serving requires online feature access with a documented freshness and lookup contract.
C:That underlying principle 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. The current incident instead turns on low-latency serving requires online feature access with a documented freshness and lookup contract.
Question 584
In an enterprise ML workflow, the reported issue is: 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 technical distinction is essential to this investigation?
- Client integrations should parse the documented model output contract instead of assuming a scalar response.
- Custom Python-model artifacts need portable packaged dependencies rather than absolute developer-local file paths.
- Custom model input contracts should validate missing and malformed values before feature transformation.
- Endpoint create/update calls and operational readiness are separate lifecycle checkpoints.
Explanation
Correct Answer
D:The technical constraint is specific to the reported failure. Endpoint create/update calls and operational readiness are separate lifecycle checkpoints. To test it, inspect deployment operation status, served entity health and known inference fixture after API completion.
Incorrect Answers
A:That underlying principle 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 current incident instead turns on endpoint create/update calls and operational readiness are separate lifecycle checkpoints.
B:That underlying principle 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. The current incident instead turns on endpoint create/update calls and operational readiness are separate lifecycle checkpoints.
C:That underlying principle 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 current incident instead turns on endpoint create/update calls and operational readiness are separate lifecycle checkpoints.
Question 585
Review this production model incident: 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. What is the central engineering constraint behind this behavior?
- Time-series data profiles use a timestamp and granularity to compare measurements across periods.
- A fixed baseline must be versioned and immutable for the intended comparison period.
- Time-series profile configuration needs a valid temporal column with reliable timestamp semantics.
- Observability starts with traceable, appropriately governed prediction logging before statistical analysis.
Explanation
Correct Answer
C:The technical constraint is specific to the reported failure. Time-series profile configuration needs a valid temporal column with reliable timestamp semantics. To test it, test timestamp conversion and timezone interpretation for valid and malformed synthetic source rows.
Incorrect Answers
A: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 time-series profile configuration needs a valid temporal column with reliable timestamp semantics.
B: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 time-series profile configuration needs a valid temporal column with reliable timestamp semantics.
D:That underlying principle is directed at a different drift detection and lakehouse monitoring situation: A serving endpoint appears healthy but nobody records model version, request identifier or prediction value, so a drift incident cannot be reconstructed. The current incident instead turns on time-series profile configuration needs a valid temporal column with reliable timestamp semantics.
Question 586
Consider the following fictional Databricks production case: An Optuna search runs 80 hyperparameter trials but records only the last winning score in MLflow, making comparisons impossible. Which implementation is most appropriate for the stated incident?
- Connect supported Optuna study execution to MLflow per-trial tracking and preserve the selected objective value.
- Select a suitable larger device or validated model-parallel/sharded execution design before scaling replicas.
- Use an appropriate Ray actor-based framework while reserving Spark for eligible data preprocessing.
- Use a supported distributed-training framework with explicit synchronization or adopt independent per-partition models only if the problem allows it.
Explanation
Correct Answer
A:This corrects the relevant failure boundary instead of merely hiding the symptom. Connect supported Optuna study execution to MLflow per-trial tracking and preserve the selected objective value. The issue matters because the winning configuration cannot be reproduced or independently audited.
Incorrect Answers
B:That corrective action is directed at a different scaling and tuning situation: An image model needs 90 GB accelerator memory for one training replica while each available GPU has 24 GB and engineers propose only adding more identical nodes. The relevant corrective action here is connect supported Optuna study execution to MLflow per-trial tracking and preserve the selected objective value.
C:That corrective action is directed at a different scaling and tuning situation: A reinforcement-learning simulator requires stateful actors exchanging messages and irregular task graphs, but the team plans one Spark SQL groupBy per simulation step. The relevant corrective action here is connect supported Optuna study execution to MLflow per-trial tracking and preserve the selected objective value.
D:That corrective action 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. The relevant corrective action here is connect supported Optuna study execution to MLflow per-trial tracking and preserve the selected objective value.
Question 587
The engineering team encounters the following issue: A team already has a vectorized Spark DataFrame pipeline and proposes converting everything to Ray datasets solely to perform a simple SQL aggregation. What direct consequence should the owner prioritize if no fix is made?
- Unnecessary system transitions add serialization and operational maintenance without model benefit.
- The apparent improvement is a data-split artifact rather than a better model.
- A single skewed group repeatedly kills workers despite healthy median group runtime.
- Available accelerators sit idle while the tuning window expires.
Explanation
Correct Answer
A:This consequence follows from the specific unresolved condition. Unnecessary system transitions add serialization and operational maintenance without model benefit. The causal constraint is: Spark SQL is often the more direct choice for distributed tabular transformations, while Ray targets flexible task/actor workloads.
Incorrect Answers
B:That direct consequence 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 more immediate risk here is unnecessary system transitions add serialization and operational maintenance without model benefit.
C:That direct consequence is directed at a different scaling and tuning situation: One enterprise account holds 80% of historical events, and grouped pandas training for that account exceeds executor memory while other accounts finish. The more immediate risk here is unnecessary system transitions add serialization and operational maintenance without model benefit.
D:That direct consequence is directed at a different scaling and tuning situation: A team runs independent GPU-heavy neural trials sequentially, despite having several isolated GPU workers that can fit each model. The more immediate risk here is unnecessary system transitions add serialization and operational maintenance without model benefit.
Question 588
The engineering team encounters the following issue: A monitoring dashboard reports a statistically significant tiny difference after one million observations, and the team immediately treats it as operationally material drift. What investigation would most effectively isolate this behavior?
- Inspect received alert fields and the corresponding monitor query, model tag and comparison window.
- Inspect sample size, numeric drift statistic and estimated business impact for the same comparison.
- Inspect the stored snapshot rows and timestamps to determine whether historical hourly observations are retained.
- Inspect endpoint status codes, resource saturation and model-load logs together with current profile metrics.
Explanation
Correct Answer
B:This is the most discriminating check for the observed symptoms. Inspect sample size, numeric drift statistic and estimated business impact for the same comparison. The underlying technical issue is that statistical significance and practical effect size differ; large samples can make negligible shifts detectable.
Incorrect Answers
A:That diagnostic evidence 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. For the stated incident, the decisive observation is inspect sample size, numeric drift statistic and estimated business impact for the same comparison.
C:That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A retail profile is created on a static materialized daily inventory table; a team expects its initial snapshot to show each prior hour as a separate trend. For the stated incident, the decisive observation is inspect sample size, numeric drift statistic and estimated business impact for the same comparison.
D:That diagnostic evidence 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. For the stated incident, the decisive observation is inspect sample size, numeric drift statistic and estimated business impact for the same comparison.
Question 589
The engineering team encounters the following issue: A multilingual prediction service must handle regional data residency, but the team copies a shared feature cache across regions without governance. What direct consequence should the owner prioritize if no fix is made?
- A test release mutates production model aliases despite otherwise successful deployment.
- Sensitive records cross an unapproved residency boundary through convenience caching.
- A release fails when a renamed or unsupported field is assumed to work unchanged.
- Manual drift creates unreproducible endpoint permissions or traffic settings.
Explanation
Correct Answer
B:This consequence follows from the specific unresolved condition. Sensitive records cross an unapproved residency boundary through convenience caching. The causal constraint is: Scalable environment design must account for data locality, residency and endpoint boundaries.
Incorrect Answers
A:That direct consequence is directed at a different environment architectures situation: A bundle references a model registry name without qualifying its environment, and a development deployment accidentally updates production registered-model aliases. The more immediate risk here is sensitive records cross an unapproved residency boundary through convenience caching.
C:That direct consequence is directed at a different environment architectures situation: An ML platform moves its configuration from older Databricks Asset Bundles terminology to newer Declarative Automation Bundles references and assumes every resource schema stayed identical. The more immediate risk here is sensitive records cross an unapproved residency boundary through convenience caching.
D:That direct consequence is directed at a different environment architectures situation: A model serving endpoint is created manually in the UI and every staging deployment differs subtly from production because no deployment definition is saved. The more immediate risk here is sensitive records cross an unapproved residency boundary through convenience caching.
Question 590
A production ML review contains this situation: A client invokes a registered custom ML model using an unauthenticated HTTP POST to the serving endpoint and assumes network access implies authorization. Which check shows that this particular problem has been resolved?
- An endpoint reaches a ready state and a test prediction succeeds before the pipeline marks the release complete.
- A permitted test identity receives a score while an unapproved identity gets access denied.
- Boundary requests return validated predictions or controlled informative errors according to contract.
- A newly registered version is discoverable under the correct catalog/schema with its tracked source run.
Explanation
Correct Answer
B:This verification measures the required result rather than just process liveness. A permitted test identity receives a score while an unapproved identity gets access denied. It confirms whether the targeted action succeeded: use a properly scoped authorized credential and supported authenticated inference request.
Incorrect Answers
A:That acceptance test is directed at a different custom model serving 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. Here the actual success criterion is a permitted test identity receives a score while an unapproved identity gets access denied.
C:That acceptance test 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. Here the actual success criterion is a permitted test identity receives a score while an unapproved identity gets access denied.
D:That acceptance test is directed at a different custom model serving situation: A research group wants to register a custom classification model in Unity Catalog but uses a local path as its complete model identity and has no governed model name. Here the actual success criterion is a permitted test identity receives a score while an unapproved identity gets access denied.