HikeLetter | PMP Training & Certification Online Course | Project Management Learning

1

For a customer-risk prediction service, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

A. Overfitting, leakage, and whether the validation split reflects the deployment scenario B. The notebook's display theme C. The model's file name D. Whether the experiment has a long title

2

When reviewing a model candidate, a data scientist needs a reproducible feature transformation applied consistently in training and serving. What is a strong design choice?

A. Use a shared, versioned transformation pipeline and test training-serving parity B. Manually implement different transformations in each environment C. Remove feature definitions from version control D. Change transformations without rerunning evaluation

3

In an enterprise model lifecycle, a model is registered for controlled promotion to production. Why preserve versioned artifacts and lineage?

A. To support reproducibility, auditability, comparison, and safe rollback B. To guarantee every model is unbiased C. To eliminate the need for validation D. To avoid documenting data sources

4

For a model deployed across multiple regions, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

A. Overfitting, leakage, and whether the validation split reflects the deployment scenario B. The notebook's display theme C. The model's file name D. Whether the experiment has a long title

5

When preparing a release for production, a data scientist needs a reproducible feature transformation applied consistently in training and serving. What is a strong design choice?

A. Use a shared, versioned transformation pipeline and test training-serving parity B. Manually implement different transformations in each environment C. Remove feature definitions from version control D. Change transformations without rerunning evaluation

6

For a customer-risk prediction service, a model is registered for controlled promotion to production. Why preserve versioned artifacts and lineage?

A. To support reproducibility, auditability, comparison, and safe rollback B. To guarantee every model is unbiased C. To eliminate the need for validation D. To avoid documenting data sources

7

When reviewing a model candidate, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

A. Overfitting, leakage, and whether the validation split reflects the deployment scenario B. The notebook's display theme C. The model's file name D. Whether the experiment has a long title

8

In an enterprise model lifecycle, a data scientist needs a reproducible feature transformation applied consistently in training and serving. What is a strong design choice?

A. Use a shared, versioned transformation pipeline and test training-serving parity B. Manually implement different transformations in each environment C. Remove feature definitions from version control D. Change transformations without rerunning evaluation

9

For a model deployed across multiple regions, a model is registered for controlled promotion to production. Why preserve versioned artifacts and lineage?

A. To support reproducibility, auditability, comparison, and safe rollback B. To guarantee every model is unbiased C. To eliminate the need for validation D. To avoid documenting data sources

10

When preparing a release for production, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

A. Overfitting, leakage, and whether the validation split reflects the deployment scenario B. The notebook's display theme C. The model's file name D. Whether the experiment has a long title

Databricks Certified Machine Learning Professional

Overview

Course Description

Course Questions

For a customer-risk prediction service, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

When reviewing a model candidate, a data scientist needs a reproducible feature transformation applied consistently in training and serving. What is a strong design choice?

In an enterprise model lifecycle, a model is registered for controlled promotion to production. Why preserve versioned artifacts and lineage?

For a model deployed across multiple regions, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

When preparing a release for production, a data scientist needs a reproducible feature transformation applied consistently in training and serving. What is a strong design choice?

For a customer-risk prediction service, a model is registered for controlled promotion to production. Why preserve versioned artifacts and lineage?

When reviewing a model candidate, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

In an enterprise model lifecycle, a data scientist needs a reproducible feature transformation applied consistently in training and serving. What is a strong design choice?

For a model deployed across multiple regions, a model is registered for controlled promotion to production. Why preserve versioned artifacts and lineage?

When preparing a release for production, a model has excellent training performance but significantly worse performance on a validation set. What should be investigated?

Top Exams

Hot Exams