Download free
iPhone & iPad · Free to startOn your Mac? Scan to install
DP-100Microsoft Azure Data Scientist
DP-100 · 30 days left ›
PREDICTED786 ±37✓ Exam Ready · 87% confidence
Inside the app
Nine interactive practice formats, on iPhone
Azure Mastery has nine interactive formats for exam practice. The examples below show original DP-100 practice questions in Azure Mastery on iPhone and iPad. The exam simulator uses timed sessions with the published domain weights. Microsoft's live interface and question mix may differ.
What is the minimum number of training examples required for fine-tuning in Azure OpenAI Service?
10 examplesAzure OpenAI Service requires a minimum of 10 training examples to initiate fine-tuning; while quality improves substantially with more data, the documented minimum threshold is 10.
50 examples50 examples is the recommended starting point for meaningful fine-tuning results and is above the minimum; the documented minimum is 10, not 50, examples required to begin the process.
500 examples500 examples would produce good fine-tuning results but is far above the documented minimum requirement of 10 examples needed to start fine-tuning in Azure OpenAI Service.
5,000 examples5,000 examples is a large fine-tuning dataset that supports high-quality specialisation, but it overstates the minimum by 500x; the service allows fine-tuning to begin with as few as 10 examples.
Multiple choice
An original DP-100 practice question with one correct answer. Choose an option to read its written rationale.
Exam-specific sampleExplore more question formats
Which TWO of the following are required fields in a pipeline component YAML specification? Select TWO.
commandThe command field is mandatory in a command component YAML because it defines the executable instruction the component runs, typically invoking the training or processing script with its bound inputs and outputs.
environmentThe environment field is required so the component has a defined runtime image with the Python and package dependencies its command needs to execute reproducibly across compute targets.
scheduleRecurrence is configured by creating a schedule against a published pipeline job, not within a component spec, so a timing field has no place in the component definition.
endpointServing destinations are declared in online or batch endpoint and deployment YAML; a reusable component only describes a computation step and never names where its results are hosted.
Multi-select
An original DP-100 multi-select question. Select all the correct options to earn the mark.
All-or-nothing
Arrange the steps to create and configure an Azure Machine Learning workspace using the Azure portal in the correct order.
⋮⋮1Select the subscription and resource group for the workspace
⋮⋮2Provide the workspace name, region, and associated resources
⋮⋮3Configure networking settings including public or private access
⋮⋮4Review the configuration and click Create to deploy the workspace
Drag-and-drop
An original DP-100 ordering question you can answer by touch on iPhone and iPad.
Interactive item
Select the settings required to secure an Azure ML workspace with network isolation.
Hotspot
An original DP-100 question with a visual prompt, designed for touch on iPhone and iPad.
Tap target
Fabrikam ML Platform Design Fabrikam is moving regional data science teams in North America, Europe and Asia-Pacific to Azure Machine Learning.
1Which workspace architecture should Fabrikam implement to meet Requirements 1…
2Which combination of compute and environment strategies should Fabrikam…
Case studies
An original DP-100 case study with several questions about the same requirements and environment.
Multi-question
✕Your answer: 50 examples
✨ Answer Coach:50 examples is the recommended starting point for meaningful fine-tuning results and is above the minimum; the documented minimum is 10, not 50, examples required to begin the process.
— grounded in authored certification guidance
Answer Coach
Answer Coach explains why the answer is correct and, where available, why each option is right or wrong. It helps you understand a mistake and remember the distinction. Read explanations after each question or at the end of a practice test. On supported devices, optional AI can reword a note on your device after checking it against the written guidance.
App exclusive
DP-100 Retired Exam Practice & Next Steps — Microsoft Azure Data Scientist
Review the final DP-100 outline on iPhone or iPad, preserve the Azure Machine Learning skills you learned, and compare them with current Microsoft credential options. Core study stays on-device and works offline; optional sync uses your private iCloud account. See every retired Microsoft exam and its successor on the retired exams hub.
The exam
What is the DP-100 exam?
DP-100 was the exam for Microsoft Certified: Azure Data Scientist Associate. Microsoft retired both the exam and certification on 1 June 2026, so DP-100 can no longer be scheduled. AI-300 is a current adjacent option for operationalising machine learning and generative AI solutions, but it is not a continuation of the retired credential under the same name. This page and question bank remain available for reference.
DP-100 is hands-on and code-aware. It validates that you can use the Azure Machine Learning workspace end-to-end — designing data assets and compute, exploring data with notebooks and AutoML, training models with MLflow tracking and Sweep jobs for hyperparameter tuning, deploying to managed online and batch endpoints, and now optimising language models with prompt engineering, fine-tuning, and RAG patterns. Expect scenario questions that show you a YAML pipeline spec or Python SDK snippet and ask you to predict behaviour — not just describe an algorithm.
Microsoft last updated the DP-100 skills outline on 11 April 2025, adding the "Optimize language models for AI applications" domain at 25–30%. Every question in Azure Mastery's DP-100 bank is mapped to that final outline. Read the archived official outline at learn.microsoft.com.
QuestionsTypically 40–60; formats vary
Duration100 minutes
Pass score700 / 1000
CostNot schedulable
CredentialAzure Data Scientist Associate retired
StatusRetired 1 June 2026
Skills measured · April 2025
DP-100 exam objectives
Four domains, with weights set by Microsoft's April 2025 update. Every domain summary below is paraphrased from the official skills outline; Azure Mastery tags questions by exam objective, so you can see which domains need more practice.
Data skill mapSourcesTransformStoreAnalyse
Design and prepare a machine learning solution20–25%
Explore key topics
The setup phase. Covers Azure Machine Learning workspace design — assets (data, environments, models), compute targets (compute instances, compute clusters, attached compute, serverless), and identity (managed identities, role assignments). Plus data assets (URI, MLTable), datastore registration (Blob, ADLS Gen2, key-based vs identity-based access), and curated vs custom environments. Around 8–15 questions per sitting.
Explore data, and run experiments20–25%
Explore key topics
Notebook-driven exploration in compute instances or Visual Studio Code, plus the Designer low-code surface and Automated ML for the no-code path. Covers MLflow integration for experiment tracking and metric logging, hyperparameter tuning with Sweep jobs (sampling strategies, early termination policies, primary metric), and choosing between AutoML and a manual pipeline. Around 8–15 questions.
Train and deploy models25–30%
Explore key topics
Tied for the largest domain. Job configuration — command jobs, pipeline jobs, parallel jobs, run context, output handling. Model registration (MLflow vs custom format) and versioning. Deployment to managed online endpoints (real-time, blue/green and traffic split), batch endpoints, and Kubernetes-attached compute. Inference monitoring, data drift detection, and retraining pipelines. Around 10–18 questions.
Optimize language models for AI applications25–30%
Explore key topics
New in the April 2025 outline, tied for the largest domain. Covers prompt engineering basics (system prompts, few-shot examples, chain-of-thought), fine-tuning workflows on Azure Machine Learning and Azure AI Foundry, retrieval-augmented generation (RAG) with embeddings and Azure AI Search, evaluating LLM outputs (groundedness, relevance, fluency, similarity), and responsible-AI considerations specific to generative AI. Around 10–18 questions.
Designed for DP-100
How Azure Mastery preserves DP-100 coverage
Exam-specific practice
350 DP-100 practice questions aligned to the published skills outline.
Coverage follows the official four domains (design/prepare ML, explore/experiment, train/deploy, optimize LLMs).
Practise decisions from the exam outline with scenarios and interactive questions.
Predicted score
Exam IQ predicts your DP-100 score on-device after roughly 30 questions.
A confidence range shows how stable the prediction is.
See which exam objectives need the most work.
Adaptive study plan
Your study plan updates as you answer questions.
Missed topics return sooner; consistently mastered topics appear less often.
Sessions focus on the gaps most likely to help you reach the 700 pass score.
Knowledge decay
Track which topics you're starting to forget.
Get reminders to revisit those topics before exam day.
Weak-spot sessions bring fading topics back into practice.
Exam rehearsal
Practise staying focused through a full 100-minute session.
Original questions follow the published domain weighting.
Microsoft controls the live interface and exact question mix.
Answer Coach
Every answer has a written explanation, with reasons for individual options where available.
See why you missed an answer, how the options differ, and what to remember.
On supported devices, optional AI can reword notes while checking them against the written explanations.
Aura guidance
Aura suggests what to do after each session.
Recaps show your progress and which topics to practise next.
Private by design
Practice, scoring, readiness, and coaching run on-device.
You don't need an Azure Mastery account or a server to process your study data.
Optional sync uses your private iCloud account.
6-week study plan
Suggested DP-100 study plan
Use these six weeks of study blocks for historical DP-100 revision. Finish by comparing the retired outline with available routes such as AI-300.
Machine Learning workspace
Configure assets, environments, identities and roles.
Compare compute instances and clusters.
Review attached and serverless compute.
Data and environments
Compare URI and MLTable assets.
Register Blob and ADLS Gen2 datastores with suitable access.
Choose curated or custom environments.
Explore and track experiments
Use notebooks for data exploration.
Compare Designer and Automated ML.
Track metrics and artifacts with MLflow.
Hyperparameter tuning
Configure sweep jobs.
Compare grid, random and Bayesian sampling.
Choose metrics and early-termination policies.
Training jobs
Compare command, pipeline and parallel jobs.
Review run context and outputs.
Explore distributed training basics.
Models and online endpoints
Register and version MLflow or custom models.
Trace model lineage.
Review online endpoints, blue-green deployment and traffic splits.
Batch deployment
Explore batch endpoints.
Review Kubernetes-attached compute.
Match deployment modes to scenarios.
Monitor model quality
Distinguish data drift and model drift.
Plan retraining triggers.
Review fairness and explainability dashboards.
Prompt engineering
Compare prompt patterns and examples.
Review system instructions.
Evaluate prompts in Foundry.
LLM optimisation
Review fine-tuning workflows.
Explore embeddings and RAG with Azure AI Search.
Compare groundedness, relevance, fluency and similarity metrics.
Target weaker topics
Run Focus Weak Spots sessions.
Revisit training, deployment and LLM optimisation.
Review explanations for recurring mistakes.
Review and choose a route
Use historical simulator questions for reference.
Compare DP-100 with routes such as AI-300.
Choose further study before selecting a credential.
Frequently asked
DP-100 FAQs
Can I still take the DP-100 exam?
No. Microsoft retired DP-100 and the Azure Data Scientist Associate certification on 1 June 2026. The exam can no longer be scheduled.
What should I study after DP-100 retired?
Microsoft did not continue the retired Azure Data Scientist Associate credential under the same name. AI-300 is a current adjacent option focused on operationalising machine learning and generative AI solutions, and Microsoft lists AI-300 courseware as the successor to DP-100 courseware.
What happens if I already earned Azure Data Scientist Associate?
A certification earned before retirement remains on your Microsoft Learn transcript. DP-100 and its renewal assessment are retired, so the exam cannot be retaken or newly earned.
Are DP-100 practice questions still useful?
Yes, as reference material for Azure Machine Learning, training, deployment, MLflow, responsible AI, and language-model optimisation. New candidates should compare those skills with current AI-300 and other active Microsoft credentials before choosing a path.
What is the current fundamentals route after DP-100 retirement?
AI-901 is the current Azure AI Fundamentals exam and DP-900 remains the current data-fundamentals option. DP-100 content is retained for reference, but new candidates should compare those fundamentals with the active AI-300 operations credential.
Which current credentials cover adjacent skills?
AI-300 is Microsoft's replacement route for operationalizing machine learning and generative AI. AI-103 covers developing AI apps and agents with Python and Microsoft Foundry. Neither recreates the retired DP-100 data-scientist credential exactly, so choose based on whether your work centres on AI operations or application development.
Is Azure Mastery free for DP-100 prep?
DP-100 is a retired reference pack. Previously purchased access stays available. For new study, choose a current exam and start with 50+ free questions and written explanations. Pro adds full Exam IQ insights, Answer Coach, the simulator and Ask Aura Preview for supported current exams.
Does DP-100 practice work offline?
Yes. Question practice, scoring, and the readiness prediction all run on-device, so you can study on a flight or a commute with no connection. No account is required, and your study data stays on your device with optional private iCloud sync.
Is there a DP-100 practice test mode?
Yes — timed practice sessions use original, blueprint-aligned questions, published timing guidance, and supported interactive item types. Adaptive study sessions and Focus Weak Spots then target the domains where your accuracy is lowest.
Free study guides
Free guides that pair with DP-100
Where DP-100 fits
DP-100 retirement and current certification paths
DP-100 and the Azure Data Scientist Associate certification retired on 1 June 2026. Microsoft identifies AI-300 as the replacement credential, with a changed emphasis on machine-learning and generative-AI operations.
DP-100 is retained for historical reference. Current adjacent choices include AI-300 for machine-learning and generative-AI operations, AI-103 for AI apps and agents, and DP-900 for data fundamentals.
Review DP-100 or explore AI-300
Keep 350 final-outline DP-100 questions for reference, or use Azure Mastery's current AI-300 bank and adaptive study plan for a current machine-learning operations option.