AI-103 Practice Questions 2 — 50 Questions

Question 1

You are designing an agentic workflow where an orchestrator agent delegates sub-tasks to specialized worker agents (e.g., a research agent and a coding agent). Which framework or service native to Azure AI supports multi-agent orchestration patterns?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Semantic Kernel or Microsoft AutoGen. Microsoft AutoGen and Semantic Kernel are core frameworks supported within Azure AI for multi-agent orchestration and collaborative agent workflows.

Answer

Azure Data Factory pipeline authoring

Answer

Azure Stream Analytics query editor

Answer

Azure Logic Apps standard workflow designer

Question 2

You are designing a multi-agent solution in Azure AI Foundry where agents collaborate asynchronously. One agent produces an intermediate JSON payload that must be strictly validated before being consumed by the next agent. What feature should you implement?

Accepted Answer

Structured Outputs with strict JSON Schema enforcement. Structured Outputs (JSON mode with strict schema enforcement) ensures that model completions conform precisely to a defined JSON Schema.

Answer

Model temperature set to 1.0

Answer

Azure API Management XML transformation policy

Answer

Azure AI Content Safety hate speech detection

Question 3

You are designing an AI application and want to test different foundation models (e.g., GPT-4o vs. Llama 3) side-by-side using the same prompt and test dataset. Which feature in Azure AI Foundry should you use?

Accepted Answer

Side-by-side evaluation in Azure AI Foundry playground. Model evaluation and playground comparison tools in Azure AI Foundry allow developers to compare outputs across different models side-by-side.

Answer

Azure Monitor workbook performance graphs

Answer

Azure Cosmos DB cross-region replication

Answer

Azure Front Door traffic routing

Question 4

You are configuring prompt engineering practices for a generative AI application in Azure AI Foundry. Which THREE practices should you follow to ensure high output quality and reliability? (Choose three.)

Accepted Answer

Specify the desired output format explicitly (e.g., JSON schema or markdown tables).. Effective prompt engineering involves clear separation of context/instructions, few-shot examples, and specifying output formats.

Answer

Embed all system instructions inside the user prompt string without using system roles.

Answer

Rely entirely on vague, implicit instructions to give the model creative freedom.

Question 5

You are configuring a code interpreter tool for an AI agent in Azure AI Agent Service. The agent needs to analyze a CSV dataset uploaded by a user. How does the agent execute this task securely?

Accepted Answer

The agent generates Python code to analyze the CSV and executes it inside a secure, isolated sandbox environment managed by the service.. The code interpreter runs in a secure, sandboxed execution environment where it dynamically writes and executes Python code to process data and generate files.

Answer

The agent compiles the CSV data into C# binaries and runs them on the host Azure App Service instance.

Answer

The agent sends the raw CSV data to an external public third-party execution API.

Answer

The agent translates the CSV data into SQL queries and executes them directly against the primary production database.

Question 6

You are implementing a RAG solution where documents contain complex hierarchical headings. To ensure search chunks retain sufficient context about which section they belong to, what chunking strategy should you adopt?

Accepted Answer

Hierarchical (parent-child) chunking retaining section headers. Hierarchical chunking or parent-child chunking retains parent section headings and metadata within smaller searchable child chunks, preserving document context.

Answer

Fixed-size character chunking with zero overlap

Answer

Random token sampling without structural awareness

Answer

Single-chunk whole document ingestion

Question 7

You are tuning the prompt for a generative AI application in Azure AI Foundry. The model frequently hallucinates when asked complex multi-part questions. You decide to use chain-of-thought (CoT) prompting. How should you structure the prompt to implement CoT effectively?

Accepted Answer

Instruct the model to think step-by-step and show its intermediate reasoning before providing the final answer. Chain-of-thought prompting involves guiding the model to articulate intermediate reasoning steps before arriving at the final conclusion.

Answer

Set the model seed parameter to a fixed integer value for reproducibility

Answer

Reduce the maximum completion token limit to 150 tokens

Answer

Instruct the model to output its response exclusively in JSON format with strict schema validation

Question 8

When designing prompt engineering strategies for complex reasoning tasks in Azure OpenAI, which TWO techniques help reduce hallucinations and improve output reliability? (Choose TWO)

Accepted Answer

Few-shot prompting providing high-quality input-output demonstration examples. Few-shot prompting with clear examples and chain-of-thought reasoning significantly improve complex task accuracy and reduce hallucinations.

Answer

Disabling system prompts entirely to maximize model freedom

Answer

Setting the top_p parameter to 0.0 while setting temperature to 2.0

Answer

Encoding all prompt variables in hexadecimal format

Question 9

You are monitoring an Azure OpenAI deployment in Azure AI Foundry and notice an increase in HTTP 429 (Too Many Requests) errors. Your application needs to handle these rate limit spikes gracefully. What is the recommended architectural solution?

Accepted Answer

Implement exponential backoff and jitter retry logic in the application SDK. Implementing retry logic with exponential backoff and jitter in the application code handles transient rate-limiting errors (HTTP 429) gracefully.

Answer

Restart the Azure OpenAI resource from the Azure Portal

Answer

Set the model temperature parameter to 0.0

Answer

Switch the authentication method from API keys to Managed Identity

Question 10

You are configuring automatic evaluation in Azure AI Foundry using the groundedness metric. The evaluation pipeline requires a source context and an LLM-generated response. What does the groundedness metric assess?

Accepted Answer

Whether the generated response is derived exclusively from the source context without unsupported claims. The groundedness metric measures whether the model's generated answers can be verified solely from the provided source context, detecting hallucinations.

Answer

The cost per token consumed during inference

Answer

The sentiment and tone of the user prompt

Question 11

You are designing a generative AI application in Azure AI Foundry and want to categorize user queries into predefined intent classes before routing them to specific agents. Which prompt engineering technique should you apply?

Accepted Answer

Few-shot prompting with labeled classification examples. Few-shot classification prompting provides examples of inputs paired with their corresponding categories to guide the model's classification accuracy.

Answer

Zero-shot random generation

Answer

Enabling JSON output mode without schema definition

Answer

Increasing model temperature to 1.5

Question 12

You are building a generative AI application that summarizes long legal documents. Users complain that summaries omit important clauses located in the middle of the document due to the model's attention limitations over extremely long texts. What phenomenon does this represent, and how can you mitigate it?

Accepted Answer

Lost in the middle phenomenon; mitigate using map-reduce or recursive summarization patterns. The 'lost in the middle' phenomenon occurs when LLMs overlook information in the middle of long contexts. Mitigation involves chunking or map-reduce summarization patterns.

Answer

Token overflow error; mitigate by increasing Azure AI Search semantic ranker depth

Answer

Embedding drift; mitigate by retraining the text-embedding-ada-002 model

Answer

Prompt injection; mitigate by enabling Azure AI Content Safety prompt shields

Question 13

You are implementing an agentic workflow where an agent executes tools that modify corporate data. To adhere to Responsible AI principles, you must ensure a human reviewer approves sensitive tool actions before they are executed. What design pattern should you implement in Azure AI Agent Service?

Accepted Answer

Human-in-the-loop (HITL) approval workflow for tool execution. Human-in-the-loop (HITL) review pauses agent execution before tool calls that perform write or destructive actions, awaiting explicit human approval.

Answer

Fully autonomous agent loop with max iterations set to 50

Answer

Azure API Management rate limiting policy

Answer

Asynchronous event grid notification after tool execution

Question 14

You are optimizing an Azure OpenAI deployment in Azure AI Foundry for high-throughput batch processing. Which TWO strategies or configurations should you consider? (Choose TWO)

Accepted Answer

Azure OpenAI Batch API for asynchronous high-volume jobs at reduced costs. Batch API usage and provisioned throughput units (PTUs) are designed for managing high-volume enterprise throughput and batch workloads.

Answer

Disabling SSL certificate validation in the client SDK

Answer

Setting model max_tokens to 1

Answer

Routing all traffic through public developer local workstations

Question 15

When configuring vector embeddings and search indexes in Azure AI Search for RAG, which THREE factors directly impact vector search performance and recall accuracy? (Choose THREE)

Accepted Answer

Embedding model dimension size and consistency with source generation. Embedding model dimensionality, vector similarity metric (e.g., cosine, dot product), and HNSW algorithm index parameters impact search performance and recall.

Answer

Azure App Service CPU architecture (x64 vs ARM)

Answer

Configuring Azure Front Door DNS CNAME records

Question 16

You are preparing a prompt template for a summarization task. You want to provide three distinct examples of input text and expected summaries before asking the model to process the target text. What is this prompting technique called?

Accepted Answer

Few-shot prompting. Providing examples within the prompt is known as few-shot prompting. Zero-shot has no examples, and chain-of-thought involves reasoning steps.

Answer

Retrieval-augmented generation

Answer

Chain-of-thought prompting

Answer

Zero-shot prompting

Question 17

You are monitoring your Azure OpenAI deployment in Azure AI Foundry. You notice an increase in HTTP 429 status codes. What does this error indicate?

Accepted Answer

Rate limit or token quota exceeded (Too Many Requests). HTTP 429 indicates that the request rate or token usage has exceeded your allocated quota or rate limit.

Answer

Internal server error at Microsoft

Answer

Model deployment not found

Answer

Invalid API key authentication

Question 18

You are auditing a generative AI application for responsible AI compliance. Which THREE practices should you implement to mitigate risks? (Choose 3)

Accepted Answer

Configure system messages to explicitly define boundaries. Responsible AI involves monitoring, filtering, and transparency, including content filtering and human-in-the-loop validation.

Answer

Store all user PII in plain text logs

Answer

Increase the temperature of the model to 1.0

Question 19

You are managing prompt assets in Azure AI Foundry. Your team collaborates on developing, testing, and versioning system prompts. Which feature in Azure AI Foundry should you use to store and manage these prompt templates?

Accepted Answer

Prompt flow / Prompt management workspace assets. Prompt flow and prompt management tools in Azure AI Foundry allow teams to author, version, and share prompt templates collaboratively.

Answer

Azure Storage Blob lifecycle management policies

Answer

Azure App Service deployment slots

Answer

Azure Key Vault secret versions

Question 20

When designing agentic workflows in Azure AI Foundry, what are THREE key characteristics or components of agentic design compared to traditional single-turn prompt completion? (Choose THREE)

Accepted Answer

Tool use and external API invocation capabilities. Planning, tool use, and multi-step reasoning loops distinguish agentic workflows from traditional single-turn API calls.

Answer

Hardcoding every conditional execution path in procedural code

Answer

Executing single stateless prompt completions without memory