cert
cert-1
cert-2

Pass Databricks Certified Machine Learning Professional Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
Certified Machine Learning Professional Exam - Verified By Experts
Certified Machine Learning Professional Premium File

Certified Machine Learning Professional Premium File

$69.99
$76.99
  • Premium File 216 Questions & Answers. Last Update: Nov 12, 2026

Whats Included:

  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
 
$76.99
$69.99
accept 12 downloads in the last 7 days
block-screenshots
Databricks Certified Machine Learning Professional Practice Test 18

Question 170

The engineering team encounters the following issue: 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 corrective action should the engineer prefer?

  1. Bind the selected model version to each group and join predictions on a unique provenance key.
  2. Increase the training host's memory within budget or change to a supported distributed estimator.
  3. Retain SparkML for its supported distributed estimator or migrate to a validated Ray-compatible training algorithm.
  4. Adopt validated model-parallel or sharded training to fit the model and its optimizer state.

Explanation

Correct Answer

A:

This corrects the relevant failure boundary instead of merely hiding the symptom. Bind the selected model version to each group and join predictions on a unique provenance key. The issue matters because scores multiply or silently mix old and new model versions.

Incorrect Answers

B:

That corrective action is directed at a different scaling and tuning situation: 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. The relevant corrective action here is bind the selected model version to each group and join predictions on a unique provenance key.

C:

That corrective action is directed at a different scaling and tuning situation: A team switches a SparkML training stage to Ray without verifying that its chosen estimator supports Ray's distributed fit contracts. The relevant corrective action here is bind the selected model version to each group and join predictions on a unique provenance key.

D:

That corrective action is directed at a different scaling and tuning situation: A language model's weights cannot fit on one accelerator yet a team launches pure data-parallel replicas, each loading the entire model. The relevant corrective action here is bind the selected model version to each group and join predictions on a unique provenance key.

Question 171

Consider the following fictional Databricks production case: An integration caches predictions without recording the model version and returns yesterday's values after a controlled rollback. What underlying principle determines the most appropriate interpretation?

  1. Serving responses that are cached or audited need model-version provenance and appropriate expiration rules.
  2. MLflow custom-model portability depends on the declared runtime dependencies as well as artifact bytes.
  3. Custom model input contracts should validate missing and malformed values before feature transformation.
  4. Repeatable endpoint deployment should preserve a controlled configuration definition and registered model identity.

Explanation

Correct Answer

A:

The technical constraint is specific to the reported failure. Serving responses that are cached or audited need model-version provenance and appropriate expiration rules. To test it, compare response creation time, endpoint model version and cache keys before and after rollback.

Incorrect Answers

B:

That underlying principle is directed at a different custom model serving situation: A team logs custom MLflow model weights but forgets to include required Python package versions in the model environment. The current incident instead turns on serving responses that are cached or audited need model-version provenance and appropriate expiration rules.

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 serving responses that are cached or audited need model-version provenance and appropriate expiration rules.

D:

That underlying principle 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. The current incident instead turns on serving responses that are cached or audited need model-version provenance and appropriate expiration rules.

Question 172

Consider the following fictional Databricks production case: A bank requires an immediate rollback after deployment and wants the prior serving environment intact until the new environment passes tests. Which implementation is most appropriate for the stated incident?

  1. Prepare a complete tested green environment and switch traffic after approval while retaining blue for rollback.
  2. Gate release completion on ready healthy serving entities and validated predictions.
  3. Add a version-compatible request adapter or coordinate client migration with explicit acceptance gates.
  4. Reduce candidate routing or restore champion traffic using the supported endpoint configuration rollback.

Explanation

Correct Answer

A:

This corrects the relevant failure boundary instead of merely hiding the symptom. Prepare a complete tested green environment and switch traffic after approval while retaining blue for rollback. The issue matters because rebuilding the old model under outage pressure extends an avoidable recovery window.

Incorrect Answers

B:

That corrective action 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. The relevant corrective action here is prepare a complete tested green environment and switch traffic after approval while retaining blue for rollback.

C:

That corrective action 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 relevant corrective action here is prepare a complete tested green environment and switch traffic after approval while retaining blue for rollback.

D:

That corrective action 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 relevant corrective action here is prepare a complete tested green environment and switch traffic after approval while retaining blue for rollback.

Question 173

In an enterprise ML workflow, the reported issue is: 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. Which change addresses the real cause without undermining other requirements?

  1. Log permitted request/prediction metadata to a governed inference table with retention and redaction controls.
  2. Calculate slice-specific error rates and a documented derived disparity metric with privacy safeguards.
  3. Align baseline and current cohort definitions before computing temporal drift.
  4. Include governed dataset/model identifiers, metric values and comparison periods in the alert payload.

Explanation

Correct Answer

C:

This corrects the relevant failure boundary instead of merely hiding the symptom. Align baseline and current cohort definitions before computing temporal drift. The issue matters because selection bias looks like model input deterioration even though each subgroup is stable.

Incorrect Answers

A:

That corrective action 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 relevant corrective action here is align baseline and current cohort definitions before computing temporal drift.

B:

That corrective action is directed at a different drift detection and lakehouse monitoring situation: A fairness reviewer needs the difference in false-negative rates between two approved cohorts, but the monitor computes an aggregate without cohort labels. The relevant corrective action here is align baseline and current cohort definitions before computing temporal drift.

D:

That corrective action is directed at a different drift detection and lakehouse monitoring situation: A drift alert fires but its message contains only the word anomaly without dataset, model version or baseline window. The relevant corrective action here is align baseline and current cohort definitions before computing temporal drift.

Question 174

Consider the following fictional Databricks production case: A batch scoring service loads a classifier version from last month while using today's new feature indexer, silently changing categorical positions. Which observation most directly tests the suspected failure boundary?

  1. Inspect model class and stored coefficients against the fit-versus-transform calls in the scoring job.
  2. Read the planned stage input/output dependencies and show the schema at the first failing stage.
  3. Inspect evaluator.isLargerBetter or equivalent metric semantics alongside the ordered candidate RMSE values.
  4. Compare saved preprocessing stage metadata and feature vector order across the two deployed artifacts.

Explanation

Correct Answer

D:

This is the most discriminating check for the observed symptoms. Compare saved preprocessing stage metadata and feature vector order across the two deployed artifacts. The underlying technical issue is that scoring should use preprocessing components trained and versioned with the exact estimator artifact.

Incorrect Answers

A:

That diagnostic evidence is directed at a different using spark ml situation: An analyst loads a fitted LogisticRegressionModel but calls fit on it for every scored batch rather than transform. For the stated incident, the decisive observation is compare saved preprocessing stage metadata and feature vector order across the two deployed artifacts.

B:

That diagnostic evidence 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. For the stated incident, the decisive observation is compare saved preprocessing stage metadata and feature vector order across the two deployed artifacts.

C:

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 compare saved preprocessing stage metadata and feature vector order across the two deployed artifacts.

Question 175

A production ML review contains this situation: An alert query uses one global transaction-failure count and masks a complete outage in a small high-value region. Which diagnostic check provides the strongest direct evidence?

  1. Inspect inference-record completeness for request, model version, inputs, outputs and timestamps.
  2. Compare labeled performance for new and established users at the same evaluation horizon.
  3. Compare regional event counts and failure proportions against the aggregate alert trigger.
  4. Compare row-level model-version IDs and profile grouping behavior across known rollout periods.

Explanation

Correct Answer

C:

This is the most discriminating check for the observed symptoms. Compare regional event counts and failure proportions against the aggregate alert trigger. The underlying technical issue is that global thresholds can hide local critical slices; alerting should reflect operational segmentation where justified.

Incorrect Answers

A:

That diagnostic evidence 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. For the stated incident, the decisive observation is compare regional event counts and failure proportions against the aggregate alert trigger.

B:

That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A recommender reports stable overall accuracy while click-through prediction quality for first-time users deteriorates sharply. For the stated incident, the decisive observation is compare regional event counts and failure proportions against the aggregate alert trigger.

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 compare regional event counts and failure proportions against the aggregate alert trigger.

Question 176

A production ML review contains this situation: An ML project has a single sequential external feature-service dependency and expects data-parallel workers to increase throughput 20-fold. What is the most material immediate risk in this situation?

  1. Every replica hits device OOM even though aggregate cluster memory appears ample.
  2. Multiple trials contend for the same accelerator and fail with memory exhaustion.
  3. More workers only saturate the service, increasing retries and overall latency.
  4. Trials are lost or repeated even when compute nodes appear healthy.

Explanation

Correct Answer

C:

This consequence follows from the specific unresolved condition. More workers only saturate the service, increasing retries and overall latency. The causal constraint is: Parallelizable fraction and upstream service limits bound the benefits of adding model-training workers.

Incorrect Answers

A:

That direct consequence 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 more immediate risk here is more workers only saturate the service, increasing retries and overall latency.

B:

That direct consequence is directed at a different scaling and tuning situation: A Ray Tune training function requests zero GPU resources but initializes a large GPU model inside every trial. The more immediate risk here is more workers only saturate the service, increasing retries and overall latency.

D:

That direct consequence 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. The more immediate risk here is more workers only saturate the service, increasing retries and overall latency.

Question 177

The engineering team encounters the following issue: An ML model performs well on routine orders but has a severe error spike on high-value shipments that comprise just 2% of volume. What direct consequence should the owner prioritize if no fix is made?

  1. The metric fails or silently excludes malformed observations, distorting the reported trend.
  2. Users receive no predictions even though the registered model metrics look excellent.
  3. The organization loses disproportionately on high-value failures hidden by average accuracy.
  4. Routine weekend traffic launches unnecessary retraining and distracts engineers.

Explanation

Correct Answer

C:

This consequence follows from the specific unresolved condition. The organization loses disproportionately on high-value failures hidden by average accuracy. The causal constraint is: Business-impact slices should capture rare high-cost cases rather than only high-count overall segments.

Incorrect Answers

A:

That direct consequence is directed at a different drift detection and lakehouse monitoring situation: An ML owner adds a custom metric computed from the average of a string-typed currency column without validating conversion. The more immediate risk here is the organization loses disproportionately on high-value failures hidden by average accuracy.

B:

That direct consequence is directed at a different drift detection and lakehouse monitoring situation: A prediction endpoint returns HTTP 503 for intermittent bursts while the model's offline accuracy remains unchanged. The more immediate risk here is the organization loses disproportionately on high-value failures hidden by average accuracy.

D:

That direct consequence 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. The more immediate risk here is the organization loses disproportionately on high-value failures hidden by average accuracy.

Question 178

Review this production model incident: 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. Which mitigation resolves the defect rather than concealing it?

  1. Route availability errors to serving reliability remediation, not to automatic training.
  2. Aggregate or minimize the custom metric and apply authorized monitoring-table permissions.
  3. Retain a stable reference comparison in addition to short-term change detection.
  4. Join labels to their originating predictions using stable IDs and a defined maturity window.

Explanation

Correct Answer

C:

This corrects the relevant failure boundary instead of merely hiding the symptom. Retain a stable reference comparison in addition to short-term change detection. The issue matters because gradual cumulative drift becomes invisible despite material divergence from the trained population.

Incorrect Answers

A:

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 retain a stable reference comparison in addition to short-term change detection.

B:

That corrective action 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. The relevant corrective action here is retain a stable reference comparison in addition to short-term change detection.

D:

That corrective action is directed at a different drift detection and lakehouse monitoring situation: A model performance dashboard compares this week's labels to last week's predictions without joining by stable request ID or consistent outcome horizon. The relevant corrective action here is retain a stable reference comparison in addition to short-term change detection.

Question 179

Consider the following fictional Databricks production case: A business deploys a model whose features are computed from a stream in memory only and disappear when its compute cluster restarts. Which observation most directly tests the suspected failure boundary?

  1. Compare last available reading age with each prediction cutoff and its permitted freshness window.
  2. Inspect state store durability, output table persistence and recovery behavior under restart.
  3. Inspect feature refresh run metadata, table versions and transformation code revisions.
  4. Map feature consumers, source commit IDs and before/after results for a fixed regression fixture.

Explanation

Correct Answer

B:

This is the most discriminating check for the observed symptoms. Inspect state store durability, output table persistence and recovery behavior under restart. The underlying technical issue is that production streaming features need durable governed storage or reproducible replay, not ephemeral worker variables.

Incorrect Answers

A:

That diagnostic evidence is directed at a different advanced feature store concepts situation: A sensor classifier silently reuses features from a year-old device after months without readings, despite a maximum allowed lookback of seven days. For the stated incident, the decisive observation is inspect state store durability, output table persistence and recovery behavior under restart.

C:

That diagnostic evidence 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. For the stated incident, the decisive observation is inspect state store durability, output table persistence and recovery behavior under restart.

D:

That diagnostic evidence is directed at a different advanced feature store concepts situation: A team updates a shared on-demand feature function without version control and unexpectedly changes both batch retraining and online serving behavior. For the stated incident, the decisive observation is inspect state store durability, output table persistence and recovery behavior under restart.

Question 180

Review this production model incident: 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 failure risk follows most closely from the unresolved behavior?

  1. Scores multiply or silently mix old and new model versions.
  2. Unnecessary model partitioning adds network latency with little memory benefit.
  3. Training compute is wasted and selected parameters are biased toward only the new subset of trials.
  4. Training stalls on memory pressure even while unrelated cluster nodes are underutilized.

Explanation

Correct Answer

B:

This consequence follows from the specific unresolved condition. Unnecessary model partitioning adds network latency with little memory benefit. The causal constraint is: Data parallelism is often more natural than model-parallel networking for a small model with large independent data.

Incorrect Answers

A:

That direct consequence is directed at a different scaling and tuning situation: 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. The more immediate risk here is unnecessary model partitioning adds network latency with little memory benefit.

C:

That direct consequence 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 more immediate risk here is unnecessary model partitioning adds network latency with little memory benefit.

D:

That direct consequence is directed at a different scaling and tuning situation: 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. The more immediate risk here is unnecessary model partitioning adds network latency with little memory benefit.

Certified Machine Learning Professional Practice Tests