Databricks Machine Learning: Databricks Machine Learning Associate Practice Questions

A machine learning engineer creates a new Databricks Runtime ML cluster and wants to access Unity Catalog feature tables from it. Which statement is correct according to the current Databricks Runtime ML documentation?

1. AUnity Catalog access from Databricks Runtime ML is available only when Photon is disabled, because Photon and governed data access use incompatible execution paths on ML clusters
2. BUnity Catalog access works only on non-ML runtimes, because Databricks Runtime ML is intended for isolated experimentation and not for governed data or model workflows
3. CUnity Catalog access requires a workspace model registry first, because feature engineering and model governance must be configured in the legacy registry before catalog access is enabled
4. DTo access Unity Catalog data on Databricks Runtime ML, the compute must use Dedicated access mode, and selecting the Machine learning checkbox in the compute UI sets that access mode automatically
5. EUnity Catalog access is automatic for any Databricks Runtime ML cluster regardless of access mode, because the runtime itself grants the necessary data permissions at cluster startup time

A manager opens an MLflow run page in Databricks and wants to know what information can be inspected there. Which answer is correct?

1. AOnly the final model URI is shown there, because run parameters, metrics, source notebook details, and artifacts are available only through the Python client and not through the Databricks UI
2. BOnly notebook comments and cluster cost estimates are shown there, because MLflow run pages are meant for project collaboration notes rather than inspection of training metadata or outputs
3. COnly artifact file names are shown there, because Databricks intentionally keeps run parameters and metrics on the experiment page and does not repeat them on an individual run details page
4. DOnly the model signature is shown there, because source notebook links, run tags, and artifact navigation are reserved for Unity Catalog model pages and not exposed at the run level
5. EThe run page can show details such as run ID, parameters, metrics, tags, artifacts, and a link back to the notebook source that created the run

A data scientist has a labeled dataset but is unsure which algorithm family and hyperparameter ranges are likely to work best. She wants a Databricks capability that can produce strong baseline candidates quickly without manually coding every trial. How does AutoML help in this situation?

1. AAutoML permanently edits the source Delta tables so only statistically important columns remain, guaranteeing that all later training pipelines will use exactly the same reduced feature set without review
2. BAutoML creates a single fixed notebook that trains one default algorithm and prevents later inspection of preprocessing or tuning choices, because reproducibility requires hiding internal trial logic from users
3. CAutoML replaces feature engineering entirely by publishing every input column to an online store first, then using feature serving latency as the criterion for model and feature selection
4. DAutoML selects a winning model by comparing notebook execution time only, because Databricks treats the fastest training job as the most cost-effective proxy for generalization quality
5. EAutoML orchestrates trials across multiple algorithms and hyperparameter settings, prepares the data, ranks candidate models with evaluation metrics, and generates notebooks so the user can inspect and refine the selected approach

A regression experiment evaluates models with RMSE, and lower values are better. The engineer wants the single best active run using the MLflow API rather than the UI. Which option reflects the correct logic?

1. ACall MlflowClient.search_registered_models and order by model version number ascending, because lower version numbers in Unity Catalog correspond to earlier and therefore usually better training results
2. BSearch the experiment runs and order by the RMSE metric ascending, optionally filtering to active runs, then take the first result as the best-performing run for that objective
3. CUse MlflowClient.restore_experiment to ensure all deleted runs are visible and then select the last created run, because the newest run is the one MLflow considers best after metric logging completes
4. DUse MlflowClient.set_terminated on poorly performing runs first, because only terminated runs can participate in metric comparisons and all active runs are ignored by the search API
5. ERead the artifact directory directly from DBFS and choose the run with the largest model file, because a larger saved model generally indicates lower regression error and better feature utilization

During a custom evaluation job, an engineer saves a confusion-matrix image to a local path and wants it attached to the current MLflow run together with the trained sklearn model. What should she do?

1. AStore the image in a Unity Catalog volume only, because MLflow run artifacts cannot include local files that were generated during the same notebook session as the model training code
2. BUse MLflow only for metrics and params, and upload the image through Catalog Explorer manually, because the run UI cannot display or retain arbitrary files alongside a logged model
3. CRegister the sklearn model first and then email the confusion matrix to the team, because model registration closes the active run and prevents any additional artifact uploads to that run
4. DWithin the active run, call mlflow.log_artifact for the saved image path and use the sklearn MLflow flavor to log the trained model so both appear in the run metadata and artifacts
5. EConvert the confusion matrix into a run tag value and serialize the model as plain JSON text, because the MLflow tracking server does not support binary model files or image artifacts in the same run

A development workspace needs to compare a newly trained candidate with a production baseline model registered elsewhere, but both workspaces are attached to the same metastore. Which Unity Catalog benefit is most relevant?

1. AUnity Catalog prevents models from being visible outside the workspace where they were created, which protects production baselines from accidental comparison in development environments
2. BUnity Catalog requires every workspace to copy a model version locally before loading it, which improves reproducibility by ensuring development and production always evaluate separate physical artifacts
3. CUnity Catalog is useful only for fine-tuned LLMs, because traditional ML models still rely on the workspace registry for cross-workspace access and cannot be loaded by name from Catalog Explorer
4. DUnity Catalog mainly changes notebook artifact storage and does not materially affect model discovery, privileges, or reuse across workspaces that share the same governed metastore
5. EModels in Unity Catalog can be accessed across workspaces attached to the same metastore, making it easier to compare development candidates against governed baseline models without duplicating registry entries

A company has three Databricks workspaces for development, staging, and production. Multiple teams need to discover and reuse governed feature tables across those workspaces without recreating them locally. Why is storing feature tables in Unity Catalog generally preferable to legacy workspace-level Feature Store tables?

1. AWorkspace-level feature tables are preferable because they can always be queried from every workspace automatically, whereas Unity Catalog feature tables are intentionally hidden outside the workspace where they were first created
2. BWorkspace-level feature tables are preferable because lineage, permissions, and discovery are more granular there, while Unity Catalog limits teams to one owner and prevents cross-workspace sharing by design
3. CUnity Catalog is preferable only when a team never needs centralized governance, because account-level feature tables are harder to discover and cannot be tagged or searched in any Databricks UI
4. DUnity Catalog feature tables provide centralized governance, lineage, and cross-workspace accessibility through the metastore, making them easier to share and manage consistently than workspace-scoped feature assets
5. EThere is no practical difference between account-level and workspace-level feature tables, because both are stored with identical sharing, privilege, and model-serving behavior in current Databricks releases

A team wants full control over what is recorded during a training run rather than relying on autologging. Which pattern correctly performs manual MLflow logging?

1. ACreate a notebook experiment, fit the model, and rely on Databricks to infer any desired parameters, metrics, artifacts, and models after the run finishes, because the hosted tracking server fills in missing metadata automatically
2. BWrite metrics into a Delta table and register the trained model in Unity Catalog only, because MLflow manual logging is limited to artifacts and cannot store scalar metrics or parameters directly in a run
3. COpen an MLflow run and explicitly call APIs such as log_param, log_metric, log_artifact, and the appropriate log_model flavor so the run records chosen metadata, files, and model artifacts
4. DUse only MlflowClient.set_tag for every result, because parameters, metrics, artifacts, and models are all represented internally as tags and can be reconstructed later from a tagged run page
5. ECall create_training_set before model fitting, because manual MLflow logging is available only for feature-store-backed models and not for standard scikit-learn, XGBoost, or custom Python workflows