[NEW] Databricks Certified Machine Learning Associate

ML Workflows (19%): Orchestrating tasks using Databricks Jobs, integrating with Delta Lake for data reliability, and monitoring end-to-end pipeline health.

Model Development (31%): Data preparation, training techniques, hyper-parameter tuning, and rigorous model evaluation and selection.

Model Deployment (12%): Registering and serving models, scaling real-time inference, and tracking performance in production.

B. A summary table only, with no access to the underlying code.

C. A leaderboard of runs and a generated notebook containing the code for the best-performing model.

D. An automated script that deletes the input Delta table to save space.

E. A PDF report intended for non-technical stakeholders only.

F. A manual list of hyperparameters that the user must type into a new notebook.

Correct Answer: C

Explanation:

A (Incorrect): While it produces models, the "notebook generation" is a key feature of the Databricks implementation.

B (Incorrect): This describes "black-box" AutoML, which is the opposite of the Databricks approach.

D (Incorrect): AutoML does not delete source data.

E (Incorrect): While summaries exist, the output is highly technical and code-based.

F (Incorrect): The parameters are captured automatically via MLflow tracking.

Question 2: When using the Databricks Feature Store, what is the primary benefit of "point-in-time" lookups during model training?

B. It prevents data leakage by ensuring features are joined based on the timestamp of the observation.

C. It allows the model to predict future stock prices with 100% accuracy.

D. It automatically converts all Python code into Scala.

E. It bypasses the need for Unity Catalog permissions.

F. It speeds up the training process by ignoring historical data.

Correct Answer: B

Explanation:

A (Incorrect): Feature Store may actually increase storage needs slightly due to versioning.

C (Incorrect): No tool can guarantee 100% accuracy or predict the future perfectly.

D (Incorrect): The Feature Store does not perform code translation.

E (Incorrect): Feature Store works alongside Unity Catalog for governance; it doesn't bypass it.

F (Incorrect): Point-in-time joins actually require careful processing of historical data.

Question 3: Which MLflow component is specifically used to manage the lifecycle of a model, including transitioning it from "Staging" to "Production"?

B. MLflow Projects

C. MLflow Model Registry

D. MLflow Recipes

E. MLflow Git Integration

F. MLflow Authentication UI

Correct Answer: C

Explanation:

A (Incorrect): Tracking is for logging parameters, metrics, and artifacts during training.

B (Incorrect): Projects are a packaging format for reproducible runs.

D (Incorrect): Recipes (formerly Pipelines) are for structuring the development workflow, not managing deployment stages.

E & F (Incorrect): These are support features and not responsible for model lifecycle management.

You can retake the exams as many times as you want

This is a huge original question bank

You get support from instructors if you have questions

Each question has a detailed explanation

Mobile-compatible with the Udemy app

30-days money-back guarantee if you're not satisfied