AI-900 sample questions with answers
The other 170 questions are in the practice tool, with answers revealed as you go.
Start 20 questionsDescribe AI workloads and considerations
- A regulator asks a bank to describe, in plain terms, how its credit-scoring model reaches individual decisions. Which Responsible AI principle does the request reflect?
- A card issuer wants transactions that look unlike a customer's established spending pattern flagged while the payment is being authorised. Which workload should be used?
- Which description best matches the computer vision family of AI workloads?
- What are large language models such as GPT designed to do?
- What is the main danger of releasing an AI system without a Responsible AI governance process?
Fundamental principles of machine learning on Azure
- Clustering belongs to which learning paradigm?
- Why is a dataset divided into training and test portions?
- Which set of activities describes feature engineering?
- Which activity belongs to the deployment stage of the machine learning lifecycle?
- What does hyperparameter tuning involve?
- Why do data scientists perform feature engineering?
- A model must flag fraudulent card transactions, which are a tiny fraction of all activity. Which evaluation approach suits this problem?
- A production model's accuracy falls over three months even though monitoring shows the input feature distributions have not moved. What is the most likely explanation?
- An Azure Machine Learning designer pipeline imputes missing values with the column mean and must report trustworthy test metrics. Where should the imputation be fitted?
- A deployed endpoint's response time rose from 50 ms to 500 ms. Traffic has grown, yet CPU and GPU utilisation on the serving nodes stays low. What is the most likely cause?
- Which pair of techniques is used to guard against overfitting?
Computer vision workloads on Azure
Natural language processing workloads on Azure
- A support team wants to classify free-text customer reviews as positive, negative or neutral. Which service should it use?
- What is the core function of Azure AI Speech?
- A captioning team is choosing between two Azure AI Speech features. How does speech-to-text differ from speech translation?
- Which Azure resource is purpose-built to host and publish a conversational bot across channels such as Teams, web chat and Slack?
Generative AI workloads on Azure
- A prompt contains a task instruction and one sample of text, with no worked examples of the expected output. Which prompting pattern is this?
- Which Microsoft offering embeds generative assistance directly into Word, Excel, PowerPoint, Outlook and Teams?
- What is the purpose of abuse monitoring in Azure OpenAI Service?