Agentic AI (intermediate)plain Mode
Review all 15 questions for Agentic AI (intermediate).Plain Q&A (No XP)
#1Multiple Choice
What is the primary architectural difference between Plan-and-Execute agents and pure ReAct agents?
#2Multiple Choice
In a multi-agent Orchestrator-Worker topology, what is the role of the Orchestrator agent?
#3Multiple Choice
Why does giving a single agent 30 tools degrade performance compared to using 3 specialized multi-agent workers with 5 tools each?
#4Multiple Choice
How should an agent handle a transient 500 error returned by a web scraper tool?
#5Multiple Choice
What state management pattern prevents long-running agent threads from exceeding LLM context windows?
#6Multiple Choice
What feature in state-graph frameworks (like LangGraph) enables human approval intervention before executing high-risk nodes?
#7Scenario
You are building an autonomous GitHub PR review agent. The agent reads code changes, runs unit tests, and posts comments. What human-in-the-loop gate design should be implemented for the 'merge_pr' tool?
#8Scenario
An SRE agent encounters a tool error: 'Pod delete failed: Permission Denied'. The agent tries calling `delete_pod` 5 more times with identical arguments. How do you rewrite the system prompt to stop this behavior?
#9Scenario
Your agent needs to remember user preferences across different sessions over several weeks. Placing the entire chat history in prompt context exceeds context limits and costs $2 per call. What memory architecture solves this?
#10Scenario
You are comparing LangGraph, CrewAI, and custom Python loops for building a multi-agent system with cycles and state rollbacks. Why is a graph-based state machine better than linear chains for this use case?
#11Code Fill
Complete the LangGraph concept used to define conditional routing between graph nodes:
builder.add_conditional_edges(
'agent_node',
___,
{'continue': 'tools_node', 'end': END}
)#12Code Fill
Complete the tool parameter requirement property ensuring side-effect operations enforce idempotency keys:
{
'name': 'create_payment',
'parameters': {
'properties': {
'amount': {'type': 'number'},
'___': {'type': 'string'}
}
}
}#13Code Fill
Complete the state key in memory systems holding active conversation history threads:
class AgentState(TypedDict):
___: Annotated[list[AnyMessage], add_messages]#14Open Ended
Explain how an Orchestrator-Worker multi-agent pattern handles task delegation.
#15Open Ended