Databricks-ML-Assoc Hard Questions — 25 Difficult Questions

Question 1

A machine learning engineer is training a model on a Databricks cluster and wants the training code to run inside a container that they control, with the same Python libraries available on every node. They also want the environment recorded with the MLflow run for reproducibility. Which Databricks capability should they use?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

A Databricks container services custom Docker image specified for the cluster.. An engineer-controlled, node-consistent runtime is exactly what container services provides: a custom Docker image is used by all nodes, so library and OS versions are fixed. MLflow then records the environment alongside the run, giving both control and traceability. Cluster libraries, init scripts, and Repos each address only part of the problem without delivering an immutable image.

Answer

A Databricks Repos checkout of the training code in the workspace.

Answer

An init script that pip installs a requirements file on each node at startup.

Answer

Cluster libraries installed from PyPI scoped to the notebook.

Question 2

A machine learning engineer has an existing Databricks Model Serving endpoint named churn-endpoint serving version 3 of a model. The team has validated version 5 and wants to direct live traffic to it while keeping the deployment reversible if quality degrades. What is the most appropriate action?

Accepted Answer

Update the served entity in the endpoint configuration to reference model version 5, allowing rollback by re-pointing to version 3. The endpoint serves whichever model version URI its configuration references, so updating the served entity to version 5 redirects live traffic through a managed rollout and keeps rollback simple by re-pointing to version 3. Deleting and recreating, batch overwrites, and registry stage changes do not achieve reversible live traffic redirection.

Answer

Delete the endpoint and create a new one with a different name that serves model version 5

Answer

Transition model version 5 to Production in the Model Registry and assume the endpoint automatically follows the stage change

Answer

Keep the endpoint on version 3 and write a scheduled job that calls version 5 and overwrites predictions in the serving table

Question 3

Refer to the exhibit. What is the most likely cause of this error in a deployed MLflow model?

Accepted Answer

The client application is sending data that does not conform to the defined input schema.. The error explicitly indicates a schema mismatch between the client's payload and the model's expected input signature. This frequently occurs when downstream applications send data without proper type casting. Validating the input data before sending it to the serving endpoint is a critical step in production MLOps to prevent type-related failures and ensure robust service interaction between disparate software components in a distributed architecture.

Answer

The serving endpoint has reached its maximum memory capacity.

Answer

The model has been corrupted during the transition to Production.

Answer

The serving endpoint is not authenticated to access the data.

Question 4

Refer to the exhibit. A machine learning engineer deployed an MLflow model to Databricks Model Serving, but inference requests are failing with the error shown in the exhibit. How should the engineer resolve this issue?

Accepted Answer

Update the inference client payload to explicitly cast the 'user_age' field to an integer or long type before sending the request.. Databricks Model Serving strict signature validation enforces data types defined during model logging. When incoming payloads contain mismatched types, such as double instead of long, requests fail. The engineer must re-log the model with a correct signature or cast incoming payload data types to match the expected schema.

Answer

Increase the timeout threshold parameter in the Databricks Model Serving endpoint configuration settings.

Answer

Restart the underlying model serving worker nodes to clear stale type caches in the serving runtime container.

Answer

Switch the serving endpoint scaling policy from automatic to manual provisioning to bypass schema validation checks.

Question 5

Refer to the exhibit. A Databricks job failed to start, returning the error shown. The job depends on MLflow for tracking. What is the most likely cause of this failure?

Accepted Answer

The cluster library configuration is missing the required MLflow package.. The ClassNotFoundException indicates that the MLflow library is missing from the runtime environment of the job cluster. In Databricks, if a job is configured to use a custom or minimal environment, essential libraries must be explicitly installed. This error highlights the importance of dependency management in production ML workflows, ensuring that all necessary packages are present in the cluster environment to prevent execution failures during automated deployment or batch scoring jobs.

Answer

The Databricks Runtime version is too recent and contains a breaking change.

Answer

The MLflow tracking server is currently unreachable due to network security policies.

Answer

The job cluster has insufficient memory to load the MLflow dependency.

Question 6

Refer to the exhibit. The logs indicate a persistent connection failure for a Databricks Model Serving endpoint. What is the most likely cause?

Accepted Answer

The endpoint is overloaded or incorrectly configured.. A 'Connection refused' error coupled with a request timeout typically indicates that the serving endpoint is either overloaded, configured incorrectly with insufficient resources, or facing network connectivity issues. Since the error persists despite retries, it points to a failure in the endpoint infrastructure or scaling capacity, which requires an investigation of the endpoint's resource allocation and the network configuration within the Databricks workspace to restore service reliability.

Answer

The model artifacts are missing.

Answer

The input data format is incorrect.

Answer

The model version is archived.

Question 7

A machine learning engineer is using MLflow to track experiments. They want to compare multiple runs and identify the run that produced the best model based on a custom metric called 'weighted_f1'. They have logged this metric using mlflow.log_metric. Which MLflow UI feature allows them to sort and filter runs by this metric to quickly find the best run?

Accepted Answer

The runs table with column sorting and the filter box using metric.weighted_f1.. The MLflow UI runs table provides a sortable and filterable view of all runs in an experiment. By sorting on the 'weighted_f1' metric column or using a filter expression, the engineer can instantly locate the top-performing run. This is the standard way to compare runs and select the best model.

Answer

The experiment notes field where you can manually record the best run ID.

Answer

The model registry's stage transitions, which automatically promote the run with the highest metric.

Answer

The artifact viewer, which displays metric plots and allows sorting by value.

Question 8

A data scientist registered a model in Unity Catalog and now wants to serve it as a low-latency REST endpoint for an application. They need automatic scaling, a secure endpoint URL, and the ability to update the served model version without redeploying infrastructure. Which Databricks capability should they use?

Accepted Answer

Model Serving with a served entity pointing at the registered model version.. Model Serving provides managed, autoscaling REST endpoints backed by registered model versions. Configuring a served entity that references a Unity Catalog model and version gives a secure URL and lets the team change the served version without rebuilding the endpoint, satisfying low-latency, secure, and updatable serving needs in one capability.

Answer

A SQL warehouse with a user-defined function wrapping the model.

Answer

An all-purpose cluster running an MLflow model server manually.

Answer

A Databricks job that runs a notebook on a schedule to score requests.

Question 9

What is the primary technical limitation when deploying an MLflow model that has custom Python dependencies not included in the standard Databricks Runtime?

Accepted Answer

The serving container will fail due to missing runtime libraries.. When using custom libraries, the inference environment must be able to install and manage these dependencies. MLflow handles this by utilizing environment specifications (like Conda or pip requirements). If these dependencies are not explicitly defined or cannot be resolved at deployment time, the serving container will fail to start. This highlights the importance of properly managing the environment specification to ensure the target serving infrastructure has the necessary software components available.

Answer

The model will automatically install the missing libraries.

Answer

The model will be rejected by the registry.

Answer

The model cannot be used with MLflow.

Question 10

Which approach is most efficient for handling high-cardinality categorical features in a machine learning model while maintaining compatibility with standard Databricks model serving?

Accepted Answer

Using the Feature Store to store pre-computed embeddings.. Using Feature Store to pre-compute and transform high-cardinality features allows you to store them as embeddings or encoded vectors. This shift moves the expensive computation from the inference phase to an offline batch process, significantly reducing latency. This approach is compatible with standard model serving, as the model only receives the pre-calculated features, ensuring consistent and performant real-time inference within the Databricks ecosystem.

Answer

One-hot encoding the features inside the model inference function.

Answer

Converting features to strings and letting the model handle them natively.

Answer

Running a Spark job on every prediction request to re-encode the features.

Question 11

An ML engineer has registered a model in the Databricks Model Registry. The model must be deployed to a REST endpoint that automatically scales with traffic and provides a stable serving environment. Which Databricks capability should they use?

Accepted Answer

Databricks Model Serving with the registered model version as the served entity.. Databricks Model Serving provides a managed, serverless endpoint for registered model versions. It automatically scales with request volume and offers a stable serving environment, eliminating the need to run and maintain a web server. The other options either require manual server management or are intended for interactive or batch use, not production REST serving.

Answer

Export the model as an MLflow artifact and load it in a Databricks notebook using mlflow.pyfunc.load_model for interactive scoring.

Answer

MLflow Model Registry webhooks that trigger a deployment script when the model version transitions to Production.

Answer

A Databricks job that runs a Python web server on a cluster and exposes the model via a public URL.

Question 12

You are deploying a model using MLflow, and you want to log the custom pre-processing logic alongside the model so it is automatically applied during inference. How should you achieve this?

Accepted Answer

Use the mlflow.pyfunc.log_model() function with a custom code wrapper.. MLflow allows you to log custom Python functions or entire pipelines as part of the model using the 'pyfunc' flavor. By wrapping your pre-processing steps and model into a custom pyfunc class and logging it, you encapsulate both the transformation and the prediction logic. This is a critical best practice for production deployments, as it ensures consistent pre-processing between training and inference, eliminating training-serving skew and simplifying the deployment process for the serving endpoint.

Answer

Create a separate inference job that runs the pre-processing code before the model.

Answer

Store the pre-processing logic in a shared Delta table and call it at inference.

Answer

Log the pre-processing parameters as tags in the MLflow run and re-apply them.

Question 13

When using MLflow to track experiments, what happens if you invoke mlflow.end_run() inside a nested loop when the parent run is already active?

Accepted Answer

The active run is terminated, and subsequent logging calls outside the loop will fail if they assume the run is active.. Calling end_run() within a nested loop terminates the current active run. If it is part of a parent-child structure, ending the child prematurely could lead to incomplete data or orphaned metrics. Proper management requires careful scoping, often using context managers (with statements), to ensure that the run lifecycle is handled cleanly and that all sub-runs correctly report back to the primary experiment tracking context.

Answer

The entire experiment is deleted from the workspace.

Answer

The parent run automatically restarts to compensate for the closed child run.

Answer

The nested loop continues, but logging is redirected to the default root run.

Question 14

A machine learning engineer has a scikit-learn model registered in Unity Catalog as `prod_ml.churn.final_model` with version 3 tagged `"champion"`. The batch scoring job reads the model with `mlflow.pyfunc.spark_udf(spark, model_uri="models:/prod_ml.churn.final_model/3")`. When a data scientist registers version 4 and sets the alias `champion` to it, the batch job starts scoring with version 4 without any code change. The engineer must guarantee the batch job always scores with the reviewed version 3 until a new version is explicitly approved. What should the engineer do?

Accepted Answer

Score with `models:/prod_ml.churn.final_model/3`, and only update the job to a new version after the version is approved and the alias is moved, keeping promotion as an explicit registry operation.. Immutable numeric version URIs prevent silent model changes, while aliases are movable pointers that can drift. To ensure the job always uses the reviewed version 3 until an explicit approval, the engineer should keep the version-pinned URI and make promotion a separate registry operation. This preserves reproducibility and makes the switch to version 4 an intentional, auditable step.

Answer

Change the model URI to `models:/prod_ml.churn.final_model/3` and rely on Unity Catalog permissions to block access to all later versions.

Answer

Keep the numeric version URI `models:/prod_ml.churn.final_model/3` and require an explicit code or configuration change to move the job to a newly approved version.

Answer

Replace the model URI with `models:/prod_ml.churn.final_model@champion` so the alias resolves consistently in every environment.

Question 15

A data scientist is training a model using Databricks AutoML. They notice that the generated notebooks include a step that performs cross-validation using a specific metric. They want to change the primary metric used for model selection to area under the ROC curve (AUC). Where should they specify this in the AutoML experiment configuration?

Accepted Answer

In the AutoML experiment setup, set the 'primary_metric' parameter to 'roc_auc'.. Databricks AutoML uses the 'primary_metric' parameter to specify the metric for model selection. To use AUC, set primary_metric='roc_auc'. Other parameter names are not valid and would not achieve the desired effect.

Answer

In the AutoML experiment setup, set the 'evaluation_metric' parameter to 'roc_auc'.

Answer

In the AutoML experiment setup, set the 'optimization_metric' parameter to 'roc_auc'.

Answer

In the AutoML experiment setup, set the 'target_metric' parameter to 'roc_auc'.

Question 16

Refer to the exhibit. Why is including the `signature` and `input_example` in the `log_model` call considered a professional best practice?

Accepted Answer

It enables schema validation and improves model serving usability.. Providing a signature and input example allows MLflow to define the expected schema for the model. This metadata facilitates automatic type validation and allows the model serving infrastructure to generate accurate API documentation. This is critical for downstream consumers who need to integrate the model, as it prevents runtime integration errors and clarifies the interface expectations before the model is even deployed into production.

Answer

It forces the model to run on a GPU during inference.

Answer

It speeds up the model training process.

Answer

It reduces the storage size of the model artifact.

Question 17

An ML engineer deployed a model to a Databricks Model Serving endpoint and enabled inference tables. After a week, the team notices that some requests returned HTTP 200 but the corresponding rows in the inference table show null prediction values. They need to diagnose why predictions are missing for those requests. Which explanation is most consistent with this symptom?

Accepted Answer

The model returned a response shape that the serving layer could not map to the prediction column.. Inference tables capture request and response payloads for each served request, and the serving layer derives the prediction column from the model's response structure. When a model returns output in a format the logging layer does not recognize as a prediction, the request still succeeds with HTTP 200 but the logged prediction is null. Sampling, autoscaling, and read permissions produce different symptoms such as missing rows or errors, not present rows with null predictions.

Answer

The endpoint's autoscaling reduced replicas, causing dropped responses.

Answer

The endpoint's service principal lacked SELECT permission on the inference table.

Answer

The inference table was configured with a sampling fraction that excluded those rows.

Question 18

A data scientist is using Hyperopt with SparkTrials on Databricks to tune a model. They notice that some trials fail due to out-of-memory errors on the driver. Which adjustment should they make to improve the tuning process?

Accepted Answer

Set the SparkTrials parameter to use a smaller number of concurrent trials.. Out-of-memory errors on the driver during SparkTrials hyperparameter tuning often occur when too many trials run concurrently, overwhelming the driver's memory as it collects results. Reducing the number of concurrent trials via the SparkTrials parameter lowers the memory pressure. This is a targeted fix that balances parallelism and resource constraints.

Answer

Increase the number of Spark executors to distribute the load.

Answer

Switch to using Trials instead of SparkTrials to run trials sequentially.

Answer

Increase the driver memory by selecting a larger driver instance type for the cluster.

Question 19

A data scientist is using Databricks Feature Store to create a feature table for a recommendation model. They want to ensure that the feature table can be used for both training and batch scoring, and that it supports point-in-time correctness. Which two actions must they take when creating the feature table? (Choose two.)

Accepted Answer

Define a primary key for the feature table.. To create a feature table in Databricks Feature Store that supports training, batch scoring, and point-in-time correctness, the data scientist must define a primary key and a timestamp key. The primary key identifies entities for joins, and the timestamp key enables time-aware joins to prevent leakage. Other actions like enabling time travel manually, partitioning, or using streaming tables are not required and may be unnecessary or counterproductive. Thus, the correct choices are defining a primary key and specifying a timestamp key.

Answer

Set the table as a streaming table.

Answer

Partition the feature table by the primary key.

Answer

Enable time travel on the feature table.

Question 20

A machine learning engineer is using Databricks AutoML to train a model on a dataset with a binary target. They want to ensure the model is optimized for recall. Which AutoML configuration should they use?

Accepted Answer

Set the primary metric to recall during AutoML experiment setup. To optimize for recall, you must set recall as the primary metric in Databricks AutoML. This directs the hyperparameter tuning and model selection to maximize recall. Other options either do not change the objective or are not available in AutoML's configuration, so they would not achieve the desired optimization.

Answer

Increase the number of trials in AutoML

Answer

Manually adjust class weights in the training code

Answer

Use the default primary metric (F1 score)