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 19

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.

Question 181

Review this production model incident: A credit reporting task needs one million records scored once per week for offline account review and requests a permanently running interactive endpoint. Which mitigation resolves the defect rather than concealing it?

  1. Choose a justified supported unknown-category handling strategy and test it without retraining on confidential evaluation labels.
  2. Topologically order vectorizers, assemblers and estimators so required feature columns exist before model fitting.
  3. Run a versioned scheduled batch-scoring workflow and materialize results for downstream review.
  4. Use model.transform on new feature vectors and retrain only in an approved training workflow.

Explanation

Correct Answer

C:

This corrects the relevant failure boundary instead of merely hiding the symptom. Run a versioned scheduled batch-scoring workflow and materialize results for downstream review. The issue matters because always-on interactive capacity incurs idle serving cost without improving the offline SLA.

Incorrect Answers

A:

That corrective action 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 relevant corrective action here is run a versioned scheduled batch-scoring workflow and materialize results for downstream review.

B:

That corrective action 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 relevant corrective action here is run a versioned scheduled batch-scoring workflow and materialize results for downstream review.

D:

That corrective action 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. The relevant corrective action here is run a versioned scheduled batch-scoring workflow and materialize results for downstream review.

Question 182

The engineering team encounters the following issue: A team manually reruns a feature notebook each week before training; feature definitions change without an execution history or stable destination. Which experiment most directly demonstrates a successful fix?

  1. A fixture with three feature revisions produces exactly one eligible feature row per prediction time.
  2. A scheduled run publishes the expected governed feature-table version with traceable input versions.
  3. All registered consumer models pass feature compatibility and expected-output tests against the promoted version.
  4. Duplicated input fixtures contribute exactly once to the computed feature total.

Explanation

Correct Answer

B:

This verification measures the required result rather than just process liveness. A scheduled run publishes the expected governed feature-table version with traceable input versions. It confirms whether the targeted action succeeded: define an owned Feature Engineering Client workflow with scheduled computation and versioned feature outputs.

Incorrect Answers

A:

That acceptance test is directed at a different advanced feature store concepts situation: A training dataset combines a feature table with multiple timestamped rows per device using an equality join on device ID only. Here the actual success criterion is a scheduled run publishes the expected governed feature-table version with traceable input versions.

C:

That acceptance test 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. Here the actual success criterion is a scheduled run publishes the expected governed feature-table version with traceable input versions.

D:

That acceptance test 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. Here the actual success criterion is a scheduled run publishes the expected governed feature-table version with traceable input versions.

Question 183

The engineering team encounters the following issue: An online pricing service requires a personalized response within 70 milliseconds per request but its team schedules a full overnight batch job as the only inference route. What corrective action should the engineer prefer?

  1. Deploy a governed low-latency serving endpoint with compatible feature retrieval and load-test capacity.
  2. Call fit on appropriate training observations and transform with the resulting StringIndexerModel.
  3. Evaluate the model with class-sensitive metrics and choose thresholds aligned with false-positive and missed-fraud tradeoffs.
  4. Construct a supported streaming ingestion and scoring query with checkpointing and compatible output semantics.

Explanation

Correct Answer

A:

This corrects the relevant failure boundary instead of merely hiding the symptom. Deploy a governed low-latency serving endpoint with compatible feature retrieval and load-test capacity. The issue matters because users receive no prediction during interactive calls even though overnight scores exist.

Incorrect Answers

B:

That corrective action is directed at a different using spark ml situation: A new Spark developer calls transform on an unfitted StringIndexer instance and expects learned category labels to appear automatically. The relevant corrective action here is deploy a governed low-latency serving endpoint with compatible feature retrieval and load-test capacity.

C:

That corrective action is directed at a different using spark ml situation: A fraud detector predicts the majority no-fraud class for 99.4% of cases and advertises 99.4% accuracy while missing most positives. The relevant corrective action here is deploy a governed low-latency serving endpoint with compatible feature retrieval and load-test capacity.

D:

That corrective action is directed at a different using spark ml situation: A streaming anomaly team uses a static read of a message directory once, believing the trained SparkML model will continuously produce scores for future arriving events. The relevant corrective action here is deploy a governed low-latency serving endpoint with compatible feature retrieval and load-test capacity.

Question 184

A production ML review contains this 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. What is the most material immediate risk in this situation?

  1. A predictor is given nonexistent features or operators expect labels from a nonpredictive stage.
  2. Pipeline fitting fails with missing-column or mismatched-schema errors.
  3. The single driver can exhaust memory before training begins despite ample executor capacity.
  4. Events arriving after startup never receive anomaly scores.

Explanation

Correct Answer

B:

This consequence follows from the specific unresolved condition. Pipeline fitting fails with missing-column or mismatched-schema errors. The causal constraint is: Pipeline stages execute in declared sequence and each estimator expects columns already produced upstream.

Incorrect Answers

A:

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 pipeline fitting fails with missing-column or mismatched-schema errors.

C:

That direct consequence is directed at a different using spark ml situation: A retailer trains a demand classifier on 180 million Delta rows already partitioned across a Spark cluster, but an analyst attempts to convert the entire training set to a local pandas object. The more immediate risk here is pipeline fitting fails with missing-column or mismatched-schema errors.

D:

That direct consequence is directed at a different using spark ml situation: A streaming anomaly team uses a static read of a message directory once, believing the trained SparkML model will continuously produce scores for future arriving events. The more immediate risk here is pipeline fitting fails with missing-column or mismatched-schema errors.

Question 185

A production ML review contains this situation: A serving endpoint appears healthy but nobody records model version, request identifier or prediction value, so a drift incident cannot be reconstructed. What is the most material immediate risk in this situation?

  1. Investigators cannot tell whether a particular model version produced the anomalous scores.
  2. A broken source creates a silent feature gap while the surviving non-null distribution appears stable.
  3. On-call teams are paged repeatedly for harmless short-lived noise.
  4. The monitor fabricates apparent minute-level precision from daily totals and misleads responders.

Explanation

Correct Answer

A:

This consequence follows from the specific unresolved condition. Investigators cannot tell whether a particular model version produced the anomalous scores. The causal constraint is: Observability starts with traceable, appropriately governed prediction logging before statistical analysis.

Incorrect Answers

B:

That direct consequence is directed at a different drift detection and lakehouse monitoring situation: A nightly scoring feed drops all null-feature rows before calculating drift, making missingness jump invisible. The more immediate risk here is investigators cannot tell whether a particular model version produced the anomalous scores.

C:

That direct consequence 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. The more immediate risk here is investigators cannot tell whether a particular model version produced the anomalous scores.

D:

That direct consequence is directed at a different drift detection and lakehouse monitoring situation: A research team captures only daily batch inference totals with no row-level timestamps but insists on per-minute live inference drift analysis. The more immediate risk here is investigators cannot tell whether a particular model version produced the anomalous scores.

Question 186

In an enterprise ML workflow, the reported issue is: A fraud detector predicts the majority no-fraud class for 99.4% of cases and advertises 99.4% accuracy while missing most positives. Before modifying the pipeline, which evidence should be collected first?

  1. Compare request latency and concurrency service-level objectives with the batch schedule and observed turnaround.
  2. Inspect fitted category mappings, handleInvalid behavior and a sample newly introduced category.
  3. Inspect its declared output column and confirm the downstream estimator consumes a vector rather than predicted labels.
  4. Inspect class prevalence, the confusion matrix, minority recall and false-negative cost rather than headline accuracy.

Explanation

Correct Answer

D:

This is the most discriminating check for the observed symptoms. Inspect class prevalence, the confusion matrix, minority recall and false-negative cost rather than headline accuracy. The underlying technical issue is that accuracy alone is misleading under extreme class imbalance; minority precision, recall and PR metrics carry critical information.

Incorrect Answers

A:

That diagnostic evidence is directed at a different using spark ml situation: An online pricing service requires a personalized response within 70 milliseconds per request but its team schedules a full overnight batch job as the only inference route. For the stated incident, the decisive observation is inspect class prevalence, the confusion matrix, minority recall and false-negative cost rather than headline accuracy.

B:

That diagnostic evidence 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. For the stated incident, the decisive observation is inspect class prevalence, the confusion matrix, minority recall and false-negative cost rather than headline accuracy.

C:

That diagnostic evidence 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. For the stated incident, the decisive observation is inspect class prevalence, the confusion matrix, minority recall and false-negative cost rather than headline accuracy.

Question 187

The engineering team encounters the following issue: A customer segmentation project promotes a scoring job but omits a database migration required by its new feature schema. Which technical principle most directly explains the root of the issue?

  1. Model lifecycle deployments require compatibility checks among code, schema, feature tables and serving targets.
  2. Experiment tracking metadata is not a substitute for governed asset privileges and serving access control.
  3. Governed model registration and access controls provide traceability beyond an ad hoc personal file path.
  4. Lifecycle reproducibility spans source data, feature definitions, training configuration, model artifact and serving identity.

Explanation

Correct Answer

A:

The technical constraint is specific to the reported failure. Model lifecycle deployments require compatibility checks among code, schema, feature tables and serving targets. To test it, compare required feature schema in the release with deployed table and endpoint contracts.

Incorrect Answers

B:

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 model lifecycle deployments require compatibility checks among code, schema, feature tables and serving targets.

C:

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 model lifecycle deployments require compatibility checks among code, schema, feature tables and serving targets.

D:

That underlying principle is directed at a different model lifecycle management situation: An organization has reliable model serving but never archives feature definitions or experiment parameters and cannot reproduce predictions after drift. The current incident instead turns on model lifecycle deployments require compatibility checks among code, schema, feature tables and serving targets.

Question 188

Consider the following fictional Databricks production case: A team has accurate inference logs but its monitoring job never refreshes computed drift metrics after a table schema change. Which observation most directly tests the suspected failure boundary?

  1. Compare inference-table last-seen with metric-table last-refresh and schema errors in profiling jobs.
  2. Bucket predictions by confidence and compare matured empirical positive rates for each bucket.
  3. Check label availability, label delay and prediction-to-label matching IDs in the inference data.
  4. Compare spike duration and alert time window against the approved ten-minute persistence rule.

Explanation

Correct Answer

A:

This is the most discriminating check for the observed symptoms. Compare inference-table last-seen with metric-table last-refresh and schema errors in profiling jobs. The underlying technical issue is that monitoring pipelines require scheduled computation and output-health checks, not merely raw event storage.

Incorrect Answers

B:

That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A binary model's average predicted probability remains 0.4, but outcomes show that 0.8 predictions are now correct only half the time. For the stated incident, the decisive observation is compare inference-table last-seen with metric-table last-refresh and schema errors in profiling jobs.

C:

That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A model's inference table records output probabilities but omits ground-truth labels, yet reviewers expect immediate precision and recall metrics. For the stated incident, the decisive observation is compare inference-table last-seen with metric-table last-refresh and schema errors in profiling jobs.

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 compare inference-table last-seen with metric-table last-refresh and schema errors in profiling jobs.

Question 189

The engineering team encounters the following issue: A new model beats the champion by 0.1 percentage points on a tiny evaluation set, and the team automatically promotes it despite large uncertainty. Which technical principle most directly explains the root of the issue?

  1. Data pipeline outages must be distinguished from genuine population shift before retraining.
  2. Ground-truth delay constrains performance-triggered retraining; fresh unlabeled predictions cannot establish predictive accuracy.
  3. Small measured differences need uncertainty or minimum-evidence gates to justify risky promotion.
  4. Drift-triggered retraining needs statistically and operationally justified thresholds, not arbitrary raw metric movement.

Explanation

Correct Answer

C:

The technical constraint is specific to the reported failure. Small measured differences need uncertainty or minimum-evidence gates to justify risky promotion. To test it, inspect evaluation sample size, confidence interval or bootstrap variation around the candidate-champion delta.

Incorrect Answers

A:

That underlying principle 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 current incident instead turns on small measured differences need uncertainty or minimum-evidence gates to justify risky promotion.

B:

That underlying principle is directed at a different automated retraining situation: A credit model's true-label feedback arrives 45 days late, yet the retraining scheduler assumes yesterday's unlabeled predictions provide yesterday's observed accuracy. The current incident instead turns on small measured differences need uncertainty or minimum-evidence gates to justify risky promotion.

D:

That underlying principle 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 current incident instead turns on small measured differences need uncertainty or minimum-evidence gates to justify risky promotion.

Question 190

In an enterprise ML workflow, the reported issue is: A fraud model predicts instantly but labels arrive 21 days later; an inference dashboard treats unlabelled transactions as negative examples. Before modifying the pipeline, which evidence should be collected first?

  1. Inspect cohort sizes, allowed sensitive fields and uncertainty around the reported subgroup error rate.
  2. Test the metric with zero, small and normal baseline denominators and inspect SQL null behavior.
  3. Inspect metric reference table identity and original model training-set distribution for the deployed version.
  4. Compare inference IDs and prediction times against actual confirmed label availability.

Explanation

Correct Answer

D:

This is the most discriminating check for the observed symptoms. Compare inference IDs and prediction times against actual confirmed label availability. The underlying technical issue is that true recall and precision require mature observed labels, not fabricated negatives for pending outcomes.

Incorrect Answers

A:

That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A demographic fairness review uses a tiny sensitive subgroup with three labeled examples and reports its observed error as definitive. For the stated incident, the decisive observation is compare inference IDs and prediction times against actual confirmed label availability.

B:

That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A custom drift metric is expressed as today-minus-yesterday percentage, but yesterday's eligible count can be zero. For the stated incident, the decisive observation is compare inference IDs and prediction times against actual confirmed label availability.

C:

That diagnostic evidence is directed at a different drift detection and lakehouse monitoring situation: A model trained on last year's customer mix is monitored only against yesterday's already shifted cohort, so its original training-population mismatch is hidden. For the stated incident, the decisive observation is compare inference IDs and prediction times against actual confirmed label availability.

Certified Machine Learning Professional Practice Tests