A team wants a repeatable path from experimentation to production on Databricks. Which setup best reflects a sound MLOps strategy?
Databricks Ml Databricks Machine Learning
Question · Easy
Overfitting: I memorized the practice test
Short
Which practice most directly improves reproducibility in a Databricks MLOps workflow?
Chapter 1: Databricks Machine Learning (38% of the Exam)
1.1 MLOps Best Practices on Databricks
Study guide
What is a primary advantage of using a Databricks Runtime for Machine Learning cluster instead of a general-purpose runtime?
Tech Exam Prep Podcast
Listen on Spotify
Podcast
A data scientist needs to start prototyping quickly on Databricks with fewer library-installation steps and more predictable package compatibility. What is the best choice?
Question · Medium
How does Databricks AutoML help with model selection for a supervised tabular problem?
Which AutoML output is especially useful when a team wants to review and customize the winning approach after the automated run finishes?
1.2 Advantages of Databricks ML Runtimes and Compute Selection
A business analyst has a labeled customer-churn table and wants a strong baseline model quickly before any custom tuning. Which approach is most appropriate?
When is Databricks AutoML less suitable than a fully custom training workflow?
What is a key benefit of creating feature tables in Unity Catalog instead of keeping them only at the workspace level?
In a Unity Catalog-enabled workspace, which Python client is used to create a new feature table in Unity Catalog?
1.3 Promoting Code vs. Promoting Models
After defining a feature DataFrame with a primary key, which step actually persists the data into an existing Unity Catalog feature table?
A model should be trained with features stored in Unity Catalog. Which sequence is the most appropriate?
Which statement correctly compares online and offline feature tables?
A fraud model must look up customer features within milliseconds during each API request. Which feature-serving approach best fits?
1.4 How AutoML Facilitates Model and Feature Selection
Which information can a data scientist review in the MLflow UI for a run?
Using the MLflow Client API, what is the most direct way to identify the best run after an experiment finishes?
A user wants to manually track training results in MLflow inside a custom notebook. Which action is valid within an MLflow run?
Why do many teams prefer the Unity Catalog model registry over the workspace model registry?
1.5 AutoML Advantages in the Model Development Process
A registered model currently has alias `Champion` pointing to version 4 and alias `Challenger` pointing to version 7. To promote the challenger into production, what should the team do?
Question · Hard
You want quick descriptive statistics such as count, mean, stddev, min, and max for numeric columns in a Spark DataFrame. Which method is the most direct fit?
Databricks Ml Ml Workflows
A feature has extreme high-end values that distort model training. Which approach directly removes outliers using the IQR rule?
Which visualization is usually the clearest first choice for understanding the distribution of a single categorical feature such as `subscription_plan`?
1.6 Feature Tables in Unity Catalog vs. the Workspace Feature Store
A data scientist wants to compare two continuous features, `income` and `spend`, to see whether they move together. Which visualization is most appropriate?
A continuous feature is strongly right-skewed because a small number of users spend far more than the rest. Which simple imputation choice is usually safer than the mean for missing values?
A feature named `payment_method` contains categories such as card, ACH, and cash, with some missing values. Which basic imputation is most appropriate?
Which feature is the best candidate for one-hot encoding?
1.7 Creating a Feature Table and Writing Data to It
Why is one-hot encoding usually a poor choice for a ZIP-code column with tens of thousands of distinct values?
In which situation is a log transformation most appropriate?
A team must predict whether a customer will churn next month. The target has only two classes: churn or not churn. Which algorithm family is a natural starting point?
Databricks Ml Model Development
A model must predict exact monthly revenue values for each store. Which problem type and algorithm family best fit?
1.8 Training a Model with Features from a Feature Table
Only 5% of the training examples are fraud cases. Which technique directly addresses class imbalance during model training?
Which sampling strategy is most likely to improve minority-class learning without changing the label meaning?
In Spark ML, which statement correctly distinguishes a transformer from an estimator?
A pipeline must encode a categorical column, assemble features, and train a classifier. Which stage order is the most reasonable?
1.9 Scoring a Model with Features from a Feature Table
Why is putting preprocessing and the estimator inside one training pipeline often better than preprocessing outside the pipeline first?
When using Hyperopt `fmin`, what should the objective function return for the metric being optimized?
Which hyperparameter search method is usually the best first choice when the search space is very large and only a limited tuning budget is available?
Loading more…