Databricks Machine Learning Professional Certification Sample Questions

01. A data scientist has developed and logged a scikit-learn gradient boosting regressor model model, and then they ended their Spark session and terminated their cluster.
After starting a new cluster, they want to review the estimators_ of the original model object to analyze each of the trees in the trained model.
Which line of code can be used to restore the model object so that estimators_ is available?
a) mlflow.sklearn.load_model(model_uri)
b) client.pyfunc.load_model(model_uri)
c) mlflow.load_model(model_uri)
d) client.list_artifacts(run_id)["estimators.csv"]
e) This can only be viewed in the MLflow Experiments UI

02. Which are types of data drift?
a) Concept drift
b) Model drift
c) Feature drift
d) Data drift

03. A data scientist has made the suggestion that their team starts using Feature Store in Databricks Machine Learning. The data scientist claims that using Feature Store will meet a number of their feature management needs.
Which of the following will the team need to implement because it is not automatically provided by Feature Store?
a) Share features across workspaces
b) Measure the drift for individual features
c) Discover features used throughout the organization
d) Track where specific feature tables are used
e) Monitor the freshness of feature tables

04. A machine learning engineer has developed a decision tree model using scikit-learn, logged the model using MLflow as decision_tree_model, and stored its run ID in the run_id Python variable. They now want to deploy that model by performing batch inference on a Spark DataFrame spark_df.
Which of the following code blocks can they use to create a function called predict that they can use to complete the task?
a) predict = spark.spark_udf(
f"runs:/{run_id}/decision_tree_model"
)
b) predict = mlflow.pyfunc.spark_udf(
spark_df,
f"runs:/{run_id}/decision_tree_model"
)
c) predict = sklearn.spark_udf(
spark_df,
f"runs:/{run_id}/decision_tree_model"
)
d) predict = mlflow.pyfunc.spark_udf(
spark,
f"runs:/{run_id}/decision_tree_model"
)

05. A machine learning engineer wants to move their model version model_version for the MLflow Model Registry model model from the None stage to the Staging stage using MLflow Client client.
Which of the following code blocks can they use to accomplish the task?
a) client.transition_model_version_stage(
name=model,
version=model_version,
stage="Staging"
)
b) client.transition_model_version_stage(
name=model,
version=model_version,
stage="None"
)
c) client.transition_model_stage(
name=model,
version=model_version,
stage="Staging"
)
d) client.transition_model__stage(
name=model,
version=model_version,
from="None",
to="Staging"
)
e) client.transition_model_version_stage(
name=model,
version=model_version,
from="None",
to="Staging"
)

06. Why is it advantageous to include context in custom model classes and objects?
a) It allows for better version control of models
b) It facilitates easier integration with production environments
c) It improves the interpretability of model predictions
d) It ensures that all necessary preprocessing steps are applied consistently

07. For Advanced Experiment Tracking, which of the following are true about model signatures in MLflow?
a) They are optional but recommended
b) They define the input and output schema of the model
c) They must be manually updated with each experiment run
d) They are automatically generated for all models

08. When merging data into a Delta table, which clause are essential to specify the conditions under which a row should be updated or inserted?
a) WHEN MATCHED THEN
b) ON DUPLICATE KEY UPDATE
c) IF EXISTS THEN UPDATE ELSE INSERT
d) ON CONFLICT DO UPDATE

09. Which operations can be performed on a Feature Store table in MLflow?
a) Create
b) Overwrite
c) Merge
d) Delete

10. A data scientist has developed a model to predict whether or not it will rain using the expected temperature and expected cloud coverage.
However, the relationship between expected temperature and whether or not it rains has changed dramatically since the time period of the data on which the model was trained.
Which type of drift is present in the above scenario?
a) Label drift
b) Feature drift
c) Concept drift
d) Prediction drift
e) None of these