Databricks-ML-Pro ML Ops Practice Test 4 — 15 Questions

Question 1

An organization needs to implement a robust CI/CD strategy for their Databricks ML models. Which TWO of the following practices are recommended to ensure reliable model deployment?

Accepted Answer

Use Databricks Repos to synchronize code across development, staging, and production environments.. Implementing automated testing and environment isolation are cornerstones of MLOps. Automated unit and integration tests catch regressions early, while separating development, staging, and production workspaces prevents accidental interference with live systems. Using Git-based workflows via Databricks Repos ensures code traceability, while the Model Registry acts as the gatekeeper for model promotion, ensuring only validated models reach production environments.

Answer

Perform manual model registration directly from the production workspace notebooks.

Answer

Execute all training runs directly in the production workspace for maximum accuracy.

Answer

Store all raw data directly inside the MLflow model artifact for portability.

Question 2

Refer to the exhibit. An MLOps engineer is reviewing a JSON object representing a model in the Databricks Model Registry. The engineer wants to promote this model to the 'Production' stage using the MLflow Python API. Which command is correct?

Accepted Answer

client.transition_model_version_stage(name='revenue_forecast', version=5, stage='Production'). To transition a model version in MLflow, the `transition_model_version_stage` function is the standard method. It requires the model name, the specific version, and the target stage string. This API call is critical for programmatic CI/CD pipelines, allowing engineers to automate the promotion process based on validation results, thereby reducing manual effort and ensuring consistent deployment procedures across the organization.

Answer

client.update_model_version(name='revenue_forecast', version=5, new_stage='Production')

Answer

mlflow.register_model(model_uri='revenue_forecast/5', stage='Production')

Answer

client.set_model_stage(model='revenue_forecast', ver=5, to='Production')

Question 3

A team is building a feature store in Databricks. They need to ensure that training data and inference data are consistent to avoid training-serving skew. What is the primary benefit of using the Databricks Feature Store in this context?

Accepted Answer

It provides a unified interface for computing features, ensuring consistency between training and inference.. The Databricks Feature Store enables the reuse of feature pipelines, ensuring that the exact same transformation logic is applied during both training and inference. By decoupling feature generation from model training, it creates a 'single source of truth' for features, effectively eliminating training-serving skew and improving the reliability of models in production. This is a vital component for maintaining feature consistency across distributed teams.

Answer

It automatically scales the underlying cluster during training runs.

Answer

It forces data scientists to use only Python for feature engineering.

Answer

It eliminates the need for any data cleaning or preprocessing steps.

Question 4

When deploying a model as a real-time REST API endpoint on Databricks, which service should be used to manage the serving infrastructure, scaling, and availability?

Accepted Answer

Databricks Model Serving. Databricks Model Serving is the managed service designed specifically to deploy models as low-latency, high-availability REST endpoints. It abstracts the underlying infrastructure, providing automatic scaling, version management, and health monitoring. This service is essential for production workloads where reliability and ease of maintenance are paramount, as it removes the burden of manual server administration and infrastructure configuration from the data science team.

Answer

Databricks SQL Warehouse

Answer

Apache Spark Streaming

Answer

Databricks File System (DBFS)

Question 5

Which THREE of the following are primary responsibilities of an MLOps engineer when maintaining production ML models in Databricks?

Accepted Answer

Monitoring model performance and detecting data drift.. An MLOps engineer's role is to ensure stability, performance, and compliance. This includes monitoring model performance to detect drift, managing the CI/CD pipeline for automated deployments, and ensuring security via robust access controls. By balancing these tasks, they ensure that the machine learning system remains reliable and valuable to the business over time, effectively bridging the gap between development and operations.

Answer

Writing the core machine learning research papers for the organization.

Answer

Manually retraining every model daily regardless of performance metrics.

Question 6

What is the primary advantage of using a Model-as-Code approach in Databricks for machine learning deployments?

Accepted Answer

It enables consistent, reproducible, and versioned deployment environments.. Model-as-Code treats the entire deployment process—infrastructure, environment configuration, and code—as versioned artifacts. This allows for total reproducibility, where any production state can be rolled back or recreated using stored definitions. This practice is essential for enterprise MLOps, as it ensures that deployments are predictable, scalable, and audit-compliant, significantly reducing the risks associated with manual configuration changes in a production environment.

Answer

It ensures that the model is always trained on the latest data.

Answer

It automatically generates unit tests for all machine learning code.

Answer

It eliminates the need for any monitoring of model performance.

Question 7

Which of the following describes the purpose of a 'Gold' table in the Medallion architecture within an MLOps pipeline?

Accepted Answer

It contains validated, business-level data ready for ML model training.. The Gold layer contains highly processed, business-level data that is ready for consumption by downstream ML models or analytical applications. By ensuring data is clean, aggregated, and validated at this stage, MLOps teams can rely on high-quality features for training, which directly improves model performance and reduces the complexity of the feature engineering step in the training pipeline.

Answer

It stores raw data ingested directly from external sources.

Answer

It serves as a transient staging area for schema evolution.

Answer

It stores model artifacts and hyperparameters for versioning.

Question 8

Refer to the exhibit. An engineer is configuring a serving endpoint. Based on the configuration provided, what is the impact of the 'auto_scale' flag?

Accepted Answer

It allows the serving endpoint to adjust worker count based on request load.. The 'auto_scale' flag allows the Databricks Model Serving endpoint to dynamically adjust the number of active workers based on real-time traffic load. This ensures that the system maintains performance during peak request periods while minimizing costs during low-traffic periods. This balance of performance and efficiency is a critical aspect of managing production ML infrastructure in an enterprise Databricks environment.

Answer

It scales the number of model versions deployed in the registry.

Answer

It automatically re-trains the model when request volume drops.

Answer

It forces the endpoint to only use 10 workers permanently.

Question 9

When auditing an ML pipeline in Databricks for compliance and governance, which THREE of the following should be verified?

Accepted Answer

The Git commit hash associated with the specific training run.. Auditing requires proof of data lineage, code versioning, and access control. Verifying these components ensures that the organization can prove how a model was built, what data was used, who approved it, and that the code was properly reviewed. This level of traceability is non-negotiable for regulated industries and essential for enterprise-grade MLOps maturity.

Answer

The personal credentials of the data scientist who manually ran the training.

Answer

The raw text of every email sent between the data scientists.

Question 10

A machine learning engineer needs to track model experiments and ensure that all training parameters and metrics are captured in a reproducible way. What is the best practice for using MLflow within Databricks notebooks?

Accepted Answer

Use MLflow with 'mlflow.start_run()' and 'mlflow.autolog()' to ensure all runs are tracked automatically.. Using 'mlflow.autolog()' or explicit 'mlflow.log_params()' within a context manager ensures that every run is isolated and documented. This practice allows for easy comparison of different hyperparameter configurations, preventing data loss and providing a clear lineage from the raw data to the final model artifact, which is crucial for compliance and debugging in enterprise ML environments.

Answer

Manually log metrics to a shared Excel file stored on DBFS after the run completes.

Answer

Only log the final evaluation metric at the end of the training loop to minimize storage.

Answer

Print the parameters to the notebook output cell and rely on the notebook's revision history.

Question 11

Your organization requires that all models deployed to production undergo a drift detection check. Which approach is most effective for monitoring model performance in Databricks?

Accepted Answer

Enable Databricks Lakehouse Monitoring on the inference Delta tables to track drift automatically.. Monitoring drift requires comparing inference data distributions against the training baseline. Databricks Lakehouse Monitoring provides a managed service that automatically detects feature and prediction drift. By leveraging Delta tables as the source of truth, the monitoring service can compute statistics periodically and trigger alerts, ensuring that any degradation in model performance is identified quickly, allowing for proactive retraining or model rollback strategies.

Answer

Write a custom Spark job to compare the training table with inference logs every hour.

Answer

Configure the Model Registry to automatically retrain the model whenever performance dips below a threshold.

Answer

Rely on end-user feedback to manually flag when the model performance seems degraded.

Question 12

You are managing a Databricks environment and need to ensure that ML models are reproducible across different workspaces. Which strategy is most effective for cross-workspace model promotion?

Accepted Answer

Use a central MLflow Model Registry in Unity Catalog to share model versions across workspaces.. Using a centralized MLflow Model Registry via Unity Catalog allows models to be shared across workspaces securely. This avoids the manual export/import of artifacts and ensures that lineage, tags, and versioning remain consistent. By managing access through Unity Catalog permissions, you ensure that only authorized environments can read or promote specific models, creating a unified, compliant, and highly scalable model deployment lifecycle.

Answer

Download the model artifact from the source workspace and upload it to the target workspace UI.

Answer

Export the model as a pickled Python object and email it to the operations team.

Answer

Re-run the training notebook in each workspace to recreate the model artifact locally.

Question 13

Your company uses MLflow tracking. You want to query all experiments that achieved a specific accuracy threshold across multiple teams. What is the most efficient way to achieve this?

Accepted Answer

Use the 'mlflow.search_runs()' API to programmatically filter runs based on the accuracy metric.. The MLflow Search API (search_runs) is specifically designed to query experiments based on metrics, parameters, and tags. By using the Python MLflow client to filter runs, you can programmatically extract insights across different workspaces. This avoids manual searching, allows for automated report generation, and facilitates governance by allowing you to easily identify top-performing models or flag models that don't meet corporate performance standards.

Answer

Export all experiment metadata to a CSV and use Excel to filter for the desired accuracy.

Answer

Manually browse through each experiment folder in the MLflow UI to find the metrics.

Answer

Query the underlying Delta tables in the MLflow experiment storage location directly.

Question 14

When deploying a model to a Databricks Model Serving endpoint, how can you ensure the model scales automatically based on traffic demand?

Accepted Answer

Configure auto-scaling settings in the Model Serving endpoint definition to handle traffic fluctuations.. Databricks Model Serving utilizes auto-scaling compute resources to handle fluctuations in traffic. By configuring the serving endpoint with appropriate auto-scaling parameters, the platform dynamically adjusts the number of concurrent instances based on request load. This ensures that the application maintains low latency under heavy load while remaining cost-effective during periods of low activity, which is essential for managing production-level model inference at scale.

Answer

Manually adjust the number of instances in the endpoint configuration every time traffic increases.

Answer

Set the endpoint to a fixed number of instances that is high enough to handle peak traffic.

Answer

Use a load balancer to redirect traffic to different model endpoints depending on the time of day.

Question 15

A team is transitioning from manual experimentation to a formal MLOps pipeline. Which component should be prioritized to ensure that data used during training is reproducible?

Accepted Answer

Use Delta Lake time travel to query the training data at the specific timestamp of the model training run.. Delta Lake's time-travel capability is the standard for ensuring data reproducibility in MLOps. By using 'VERSION AS OF' or 'TIMESTAMP AS OF', you can query the exact state of the training dataset at any point in time. This is critical for auditing, model retraining, and debugging, as it eliminates the ambiguity of changes made to the underlying data sources over time.

Answer

Store all training data as CSV files in DBFS with unique filenames for each run.

Answer

Ask the data engineering team to copy the data to a private folder for every run.

Answer

Only rely on the model performance metrics to infer the quality of the training data.