Quiz for LangGraph and Agentic AI module

No.	Training Unit	Lecture	Training content	Question	Level	Mark	Answer	Answer Option A	Answer Option B	Answer Option C	Answer Option D
1	Unit 4: Multi-Agent Collaboration	Lec4	System Benefits	What is the primary advantage of “Specialization” in a multi-agent system?	Easy	1	A	Improved accuracy and efficiency by assigning domains to specialized agents.	Reduced latency by having fewer agents.	Lower cost due to using smaller system prompts.	eliminating the need for a shared state.
2	Unit 4: Multi-Agent Collaboration	Lec4	Collaboration Patterns	Which pattern describes a workflow where Agent A’s output becomes the direct input for Agent B?	Easy	1	B	Hierarchical (Supervisor)	Sequential (Pipeline)	Network (Peer-to-Peer)	Competitive
3	Unit 4: Multi-Agent Collaboration	Lec4	Hierarchical System	In a “Hierarchical” agent system, what is the primary role of the Supervisor/Primary Assistant?	Medium	1	A	To coordinate and route tasks to specialized child agents.	To execute the low-level tool logic directly.	To store the persistent history in the database.	To provide the human user interface only.
4	Unit 4: Multi-Agent Collaboration	Lec4	State Management	What is a “Dialog Stack” (or dialog_state) used for in a multi-agent LangGraph system?	Hard	1	C	To store the LLM weights for each agent.	To track the total token usage per agent.	To track the hierarchy of active agents (e.g., which agent currently has control).	To store the user’s password securely.
5	Unit 4: Multi-Agent Collaboration	Lec4	Dialog Management	In a hierarchical agentic graph, what operation is performed to return control from a child agent to the supervisor?	Medium	1	B	Push a new state.	Pop the last state from the dialog stack.	Clear the entire conversation history.	Restart the graph from the START node.
6	Unit 4: Multi-Agent Collaboration	Lec4	Context Injection	What is “Context Injection” in the context of multi-agent tool calling?	Hard	1	C	Manually typing user info into every prompt.	Hardcoding user data into the tool logic.	Automatically passing user metadata (email, ID) from the state into tool calls.	Using a vector database for context.
7	Unit 4: Multi-Agent Collaboration	Lec4	State Sharing	why is a “Shared State” critical in complex multi-agent systems?	Medium	1	D	It makes the graph linear.	It prevents the LLM from hallucinating.	It is required for the internet to work.	It allows different agents to communicate and share data/messages through a common schema.
8	Unit 4: Multi-Agent Collaboration	Lec4	Transitions	What is the purpose of an “Entry Node” when switching to a child agent?	Medium	1	A	To provide a transition message (ToolMessage) that tells the child agent to take over.	To delete the previous conversation history.	To validate the user’s login credentials.	To compile the graph for the first time.
9	Unit 4: Multi-Agent Collaboration	Lec4	Routing	How does a Supervisor typically decide which specialized agent to route to?	Medium	1	B	By random selection.	Based on the tool name requested in the Coordinator’s tool_calls.	Based on the user’s IP address.	By checking the current time.
10	Unit 4: Multi-Agent Collaboration	Lec4	P2P Pattern	What characterizes a “Network” (Peer-to-Peer) collaboration pattern?	Medium	1	C	One agent controls everyone.	Agents only work in a fixed sequence.	Agents communicate directly with each other without a central supervisor.	Agents compete to give the fastest answer.
11	Unit 4: Multi-Agent Collaboration	Lec4	Hierarchical System	In a Supervisor pattern, who manages worker routing?	Medium	1	A	A central Assistant/Supervisor.	The human user.	Each worker agent.	The database.
12	Unit 4: Multi-Agent Collaboration	Lec4	Dialog Management	What happens during a “Push State” operation?	Hard	1	B	Resetting the graph.	Adding an agent to the dialog stack.	Saving to PostgreSQL.	Deleting the last message.
13	Unit 4: Multi-Agent Collaboration	Lec4	Context Injection	Context Injection ensures agents don’t have to…	Hard	1	C	Use an LLM.	Search the web.	Manually pass user IDs.	Format JSON.
14	Unit 4: Multi-Agent Collaboration	Lec4	Routing	Why implement a “Tool Call Fallback” node?	Medium	1	D	To double cost.	To translate code.	UI purposes.	To handle failures gracefully.
15	Unit 4: Multi-Agent Collaboration	Lec4	Transitions	When should an agent use CompleteOrEscalate?	Medium	1	A	When the task is out of scope or finished.	To start a search.	To clear memory.	Every two turns.
16	Unit 4: Multi-Agent Collaboration	Lec4	State Management	Tool schemas help…	Medium	1	B	Speed up LLMs.	Provide validation and type safety.	Write Python code.	Replace docstrings.
17	Unit 4: Multi-Agent Collaboration	Lec4	State Sharing	How are results functionally shared between agents?	Hard	1	C	Email.	Global variables.	Updating the shared messages list.	Temporary files.
18	Unit 4: Multi-Agent Collaboration	Lec4	Dialog Management	Popping the dialog stack returns control to…	Medium	1	D	The END node.	A child agent.	The database.	The previous agent in hierarchy.
19	Unit 4: Multi-Agent Collaboration	Lec4	Transitions	An Entry Node creates a…	Hard	1	A	ToolMessage with conversation context.	SystemMessage only.	New graph object.	Thread ID.
20	Unit 4: Multi-Agent Collaboration	Lec4	System Benefits	ReAct agents are best for…	Easy	1	B	Complex enterprise systems.	Simple, single-domain tasks.	Department coordination.	Multi-agent teams.