Microsoft DP-100: Designing and Implementing a Data Science Solution on Azure

Last checked October 22, 2026 against the official Microsoft exam page.

1. Question 1Optimize language models for AI applications

   What is the Azure AI Foundry model catalog used for?

   1. AStoring the training data for custom models so that it can be shared across the organisation
   2. BDiscovering, comparing and deploying foundation models from many providers in one place
   3. CRecording the billing rate of each subscription so that teams can see their monthly spend
   4. DHosting the workspace notebooks so that they run without any compute attached to them at all
   Show answer and explanationHide answer

   Correct answer: B

   The model catalog lets teams discover, compare and deploy foundation models from many providers. Training data storage, billing and notebook hosting are different functions.

2. Question 2Train and deploy models

   How does a training job consume data that is registered as a data asset?

   1. AThe job reads the data from the data scientist's personal laptop each time the run starts in the cloud
   2. BThe job copies the data into the source code folder before the run, so that it is versioned with Git
   3. CThe job references the data asset as an input, and the platform mounts or downloads it at run time
   4. DThe job requires the data to be pasted into the job YAML as a long string of values for each run
   Show answer and explanationHide answer

   Correct answer: C

   Jobs reference data assets as inputs, and the platform mounts or downloads them at run time. Copying data into code folders, reading laptops and embedding values in YAML are poor practice.

3. Question 3Design and prepare a machine learning solution

   Which Azure Machine Learning compute type suits an interactive development environment used by one data scientist for notebooks?

   1. AA compute cluster, which autoscales a set of nodes for batch training jobs run by a pipeline
   2. BA compute instance, which is a managed development workstation for one user
   3. CAn inference cluster, which hosts real-time endpoints that score requests from applications
   4. DAn attached virtual machine that the team maintains by hand with its own software updates
   Show answer and explanationHide answer

   Correct answer: B

   A compute instance is a managed development workstation for an individual user. Clusters, inference clusters and hand-maintained virtual machines serve other purposes.

4. Question 4Explore data, and run experiments

   Which automated ML task type predicts a continuous numeric value from tabular data?

   1. ARegression, which predicts a continuous numeric value from the features of each row
   2. BClassification, which assigns each row to one of a fixed set of discrete labels for prediction
   3. CImage segmentation, which labels each pixel of an image with a class from a fixed set of classes
   4. DClustering, which groups unlabelled rows by similarity and has no target value to predict at all
   Show answer and explanationHide answer

   Correct answer: A

   Regression predicts continuous values from tabular features. Classification predicts discrete labels, segmentation works on images and clustering has no target.

5. Question 5Optimize language models for AI applications

   Why compare language models with benchmarks before choosing one for an application?

   1. ATo measure quality, cost and speed on tasks that resemble the application's own workload
   2. BTo choose the model with the most parameters, since size always determines the answer quality
   3. CTo find the model with the longest name, because the naming reflects the training data used
   4. DTo avoid running any test on the application's own data, which is always misleading for teams
   Show answer and explanationHide answer

   Correct answer: A

   Benchmarks plus tests on representative tasks show quality, cost and speed trade-offs. Size, naming and avoiding own-data tests are poor selection methods.