Choose the best answer for each question. Answers are at the bottom.
Q1. You want to build and test prompt flows locally. Which tool should you use? A) Azure CLI B) VS Code with Azure AI extension C) Jupyter Notebook D) REST API
Q2. Which AI framework is a Microsoft first-party SDK supporting C#, Python, and Java? A) LangChain B) LlamaIndex C) Semantic Kernel D) Haystack
Q3. You need to rate-limit Azure OpenAI calls by token count per subscription. What should you use? A) Azure Front Door B) APIM with token-limit policy C) Application Gateway D) Azure Firewall
Q4. A team wants advanced document indexing for RAG. Which framework is best suited? A) AutoGen B) Semantic Kernel C) LlamaIndex D) LangChain
Q5. You need multiple AI agents to collaborate. Which framework supports this natively? A) Semantic Kernel B) LangChain C) AutoGen D) LlamaIndex
Q6. What is semantic caching in APIM? A) Caching by exact URL match B) Caching responses for semantically similar prompts C) Caching model weights D) Caching auth tokens
Q7. You want a serverless function responding to HTTP requests with AI content. Which service? A) Container Apps B) Azure Functions C) Azure Batch D) Azure VM
Q8. A business analyst wants to auto-summarize emails with GPT-4o and post to Teams. Best low-code solution? A) Azure Functions B) Custom Python script C) Logic Apps / Power Automate D) Copilot Studio
Q9. Which CLI command deploys a model to a managed online endpoint?
A) az ai deploy create B) az ml online-deployment create C) az openai deployment create D) az ml model deploy
Q10. You’re using LangChain with Azure OpenAI. What class for chat completions?
A) OpenAI B) AzureChatOpenAI C) ChatOpenAI D) AzureOpenAI
Q11. How can you use AI Foundry endpoints with the standard openai Python package?
A) You can’t B) Set base_url to your Azure endpoint C) Install a special plugin D) Use an adapter library
Q12. Which service provides unified analytics for managing AI training data and visualizing results? A) Azure Synapse B) Azure Databricks C) Microsoft Fabric D) Azure Data Factory
Q13. Best combination for an internal Teams assistant? A) Custom web app + App Service B) Teams Bot + Copilot Studio + Azure OpenAI C) Logic Apps + Email D) Functions + REST API
Q14. APIM’s circuit breaker helps with which scenario? A) SQL injection prevention B) Failing over when backend returns 429 rate-limit responses C) Data encryption D) Schema validation
Q15. When should you choose Semantic Kernel over LangChain? A) Fastest prototyping B) Enterprise .NET apps with plugin-based AI orchestration C) RAG only D) Multi-agent conversations
| Q | Answer | Explanation |
|---|---|---|
| 1 | B | VS Code with Azure AI extension provides visual editing and local testing. |
| 2 | C | Semantic Kernel is Microsoft’s first-party SDK. |
| 3 | B | APIM’s azure-openai-token-limit policy enables token-based rate limiting. |
| 4 | C | LlamaIndex specializes in data-centric RAG with advanced indexing. |
| 5 | C | AutoGen enables multi-agent conversations and collaboration. |
| 6 | B | Semantic caching matches similar prompts using embeddings. |
| 7 | B | Azure Functions provides serverless compute with HTTP triggers. |
| 8 | C | Logic Apps/Power Automate offer no-code connectors for email, GPT-4o, and Teams. |
| 9 | B | az ml online-deployment create is the correct CLI command. |
| 10 | B | AzureChatOpenAI is the LangChain class for Azure OpenAI chat. |
| 11 | B | AI Foundry endpoints are OpenAI-compatible - just change base URL. |
| 12 | C | Microsoft Fabric unifies data engineering, analytics, and AI. |
| 13 | B | Teams Bot + Copilot Studio + Azure OpenAI is the recommended stack. |
| 14 | B | Circuit breaker detects failures and routes to healthy backends. |
| 15 | B | Semantic Kernel excels in enterprise .NET apps with plugin architecture. |