AI / Agentic AI Interview Questions
1. What is Agentic AI?
Agentic AI refers to AI systems that go beyond generating text or answering questions, they perceive their environment, reason about a goal, take real actions through tools, and adapt based on what happens, continuing this loop until a task is complete. Unlike a traditional chatbot that responds ...
2. What are the core components of an AI agent?
Most agent architectures decompose into the same handful of core components, regardless of the specific framework used to build them. Component Role Perception Handles context window management, conversation state, and input validation Reasoning Decides the next action through planning, tool sele...
3. Define the ReAct framework?
ReAct, short for Reasoning and Acting, is an agent pattern that interleaves explicit reasoning steps with tool-based actions, rather than jumping straight from input to action. Loop: Thought = LM(context, previous_observation) Action = Parse(Thought) Observation = Execute(Action) context += [Thou...
4. What is Tool Use (Function Calling) in agentic AI?
Tool use, often implemented through function calling, is how an agent extends beyond generating language and actually affects the outside world, calling APIs, databases, or code execution environments. Each tool is described with a schema, its name, its purpose, and the parameters it expects The ...
5. What are the types of memory used by AI agents?
Agent memory is generally split into a few distinct types, each serving a different purpose. Memory type What it stores Short-term memory The active conversation and context within the current session, roughly equivalent to the LLM's context window Long-term memory Persistent knowledge from past ...
6. What is a Multi-Agent System?
A Multi-Agent System splits a complex task across several specialized agents rather than relying on a single agent to handle everything itself. Each agent can focus on a narrower responsibility, such as research, coding, or review Agents can run sequentially, in parallel, or under a hierarchical ...
7. Define Chain-of-Thought reasoning?
Chain-of-Thought reasoning is a prompting technique where a model is encouraged to write out intermediate reasoning steps before arriving at a final answer, rather than jumping directly to a conclusion. Breaks a complex problem into smaller, more manageable reasoning steps Makes a model's reasoni...
8. What is the Plan-and-Execute pattern?
Plan-and-Execute is an agent architecture where the agent generates a complete plan upfront, then works through each step of that plan sequentially, rather than re-reasoning from scratch after every single action. Produces the full sequence of steps needed to reach a goal before taking the first ...
9. What is Reflexion (self-correction) in agentic AI?
Reflexion is a self-correction mechanism where an agent critiques its own recent actions and outcomes, then uses that internal feedback to adjust its future behavior within the same task. Adds an internal feedback channel on top of the normal reasoning-action loop Helps reduce compounding errors,...
10. Describe the OODA loop as applied to AI agents?
flowchart LR A[Observe] --> B[Orient] B --> C[Decide] C --> D[Act] D --> A The OODA loop, Observe, Orient, Decide, Act, is a decision-making cycle borrowed from military strategy that maps closely onto how agentic AI systems operate in practice. Observe : the agent perceives its current context a...
11. What is Retrieval-Augmented Generation (RAG)?
Retrieval-Augmented Generation grounds a language model's output in external evidence by retrieving relevant documents or data before generating a response, rather than relying purely on what the model memorized during training. A retrieval step searches an external knowledge source, often a vect...
12. What is Agentic RAG?
Agentic RAG extends traditional RAG by letting an agent decide when, what, and how many times to retrieve, rather than performing one fixed retrieval step before every generation. The agent can reason about whether the retrieved content actually answers the question, and retrieve again if it does...
13. What is the Model Context Protocol (MCP)?
The Model Context Protocol is a standardized way for AI agents and applications to share tools, resources, and context, so agents can coordinate workflows across different systems more reliably. Lets an agent discover what tools and resources are available from an MCP server, rather than needing ...
14. What are the popular frameworks used to build AI agents?
Several frameworks have emerged to help developers build agentic systems, each with a somewhat different focus. Framework Notable focus LangGraph Cyclical graph-based orchestration with explicit nodes and conditional edges LangChain General-purpose framework for chaining LLM calls, tools, and ret...
15. What is a Human-in-the-Loop checkpoint?
A Human-in-the-Loop checkpoint is a deliberate pause in an otherwise autonomous agent workflow where a human must review or approve an action before the agent proceeds. Commonly inserted before high-stakes or irreversible actions, like sending an email or executing a financial transaction Lets an...
16. Define Prompt Injection in the context of AI agents?
Prompt injection is an attack where malicious instructions are embedded in content an agent processes, tricking it into behaving differently than the user intended. Can be hidden inside a document, webpage, or file the agent reads as part of its task Exploits the fact that an LLM can't reliably d...
17. What is Context Poisoning?
Context poisoning is a more subtle variant of prompt injection where an attacker doesn't inject an obvious command, but instead subtly manipulates a document or data source an agent will later read, altering the agent's behavior indirectly. Harder to detect than a blatant injected instruction, si...
18. What is Tool Injection?
Tool injection is an attack where an attacker manipulates the instructions or metadata an agent receives about a tool, tricking it into executing a malicious function or targeting the wrong resource. Can exploit the fact that an agent typically trusts a tool's name and description without verifyi...
19. Define an Orchestrator-Worker pattern?
The Orchestrator-Worker pattern splits a multi-agent system into one coordinating agent, the orchestrator, and several specialized agents, the workers, that each handle a specific part of a larger task. The orchestrator breaks the overall goal into subtasks and assigns them to appropriate workers...
20. What is short-term memory in an AI agent?
Short-term memory in an AI agent corresponds to the active conversation and context within the current session, roughly equivalent to what fits inside the model's context window. Includes the current task, recent tool calls, and their observed results Disappears once the session ends unless it's ...
21. What is long-term memory in an AI agent?
Long-term memory lets an agent retain knowledge across sessions rather than starting fresh every time, typically implemented through an external vector database or knowledge store. Stores information as embeddings that can be searched for semantic similarity to a new query Lets an agent recall re...
22. What is episodic memory in an AI agent?
Episodic memory captures specific past events along with temporal information, letting an agent recall not just facts but what happened, and roughly when. Useful for tasks where the sequence or timing of past events matters, not just the underlying facts Distinct from general long-term factual me...
23. List the common risks associated with deploying agentic AI?
Prompt injection and context poisoning : malicious instructions hidden in content the agent processes Tool injection : manipulated tool metadata tricking the agent into harmful actions Data leakage : sensitive information exposed through tool calls or memory persistence Unauthorized persistence :...
24. What is Tree-of-Thoughts reasoning?
Tree-of-Thoughts treats reasoning as a search problem, instead of following one linear chain of thought, the model explores multiple candidate reasoning paths as branches of a tree, and can backtrack from paths that aren't working. Generates several possible next steps at each point, rather than ...
25. What is an agent's "turn budget"?
A turn budget is a limit on how many reasoning-action cycles, or turns, an agent is allowed to take before it must stop, regardless of whether it has finished the task. Prevents an agent from looping indefinitely on a task it can't complete Bounds cost and latency, since each additional turn typi...
26. Describe LangGraph's role in agent architecture?
LangGraph represents an agent's workflow as a cyclical graph of nodes and edges, rather than leaving the entire flow of control up to the language model's own judgment. Nodes represent discrete functions or steps, like drafting a response or calling a tool Edges represent the possible transitions...
27. What is Context Bloat in multi-tool agent systems?
Context bloat happens when an agent is connected to many tools or MCP servers at once, and each one injects its own tool metadata, descriptions, and prompts into the model's context window. Overwhelms the model with far more tool information than is relevant to the current task Can degrade the mo...
28. What is Grounding in the context of AI agents?
Grounding refers to connecting an agent's reasoning and language output to real, verifiable information or a real environment, rather than letting it generate plausible-sounding but unverified claims. Retrieval-augmented generation grounds answers in retrieved documents Tool use grounds actions i...
29. What is the difference between Agentic AI and a traditional chatbot?
Traditional chatbot Agentic AI Responds once per user turn Operates autonomously across multiple steps toward a goal No inherent tool use beyond generating text Calls tools, APIs, and executes actions in the real world Requires a new user prompt to continue Continues a plan-act-observe-adapt loop...
30. What is the difference between ReAct and Plan-and-Execute patterns?
ReAct Plan-and-Execute Interleaves one reasoning step with one action at a time Generates a full plan upfront, then executes steps sequentially Naturally adapts mid-task to unexpected tool results Less naturally adaptive once the plan is fixed Higher token usage due to repeated reasoning cycles T...
31. Why does ReAct carry higher latency and cost than Plan-and-Execute?
ReAct's core loop performs a full reasoning-action-observation cycle before every single step, which means every action in a multi-step task pays for another round of model inference. Each cycle consumes additional tokens for the reasoning text generated before the action Sequential reasoning and...
32. How does an AI agent use tool schemas to decide which tool to call?
An agent's reasoning engine, typically the underlying LLM, selects a tool based purely on the tool's declared name and description, without any visibility into how that tool is actually implemented. Each available tool exposes a schema describing its purpose and expected parameters The agent comp...
33. When should you use a hierarchical (cyclical graph) architecture instead of a single-loop agent?
A single-loop agent works well for tasks that are relatively linear and don't need much structural control over what happens next. Use a graph-based, hierarchical architecture when a workflow has clear branching logic, like different paths for success versus failure Useful when specific steps nee...
34. Why is a Human Approval Node inserted into sensitive agent workflows?
A Human Approval Node is a deliberate pause point in a graph-based agent workflow that blocks progress until a person reviews and approves the agent's proposed next action. Commonly placed between a draft step and an execution step, such as between drafting an email and actually sending it Protec...
35. What is the difference between RAG and Agentic RAG?
RAG Agentic RAG Performs one retrieval step before generating a response The agent decides when and how many times to retrieve Fixed pipeline: retrieve, then generate Retrieval is treated as one tool among several the agent can invoke repeatedly Struggles with multi-hop questions needing several ...
36. How does Reflexion reduce compounding errors in agent execution?
Compounding errors happen when a small mistake early in a multi-step task goes unnoticed and cascades into a badly wrong final result several steps later. Reflexion adds an internal feedback channel where the agent critiques its own recent actions and outcomes This self-critique step can catch a ...
37. Why do MCP servers remain stateless in the recommended architecture pattern?
In the common MCP architecture pattern, the MCP server acts purely as a stateless provider of tools, resources, and prompt templates, while the MCP client hosts the actual LLM runtime and agent logic that decides what to call and when. A stateless server can be reused and replaced without worryin...
38. What is the difference between MRKL-style routing and end-to-end LLM reasoning?
MRKL-style systems route tasks to specialized tools or modules, separating natural language understanding from deterministic components rather than asking the LLM to reason through everything itself. A router decides which specialized tool or module should handle a given part of the task Determin...
39. How does an OAuth-based identity flow protect an MCP-connected agent session?
Rather than letting an agent act with blanket, unverified trust, an OAuth-based flow establishes identity for the specific session, producing a token that represents a particular user acting through a particular agent, with defined scopes. Every action the agent takes carries an identifier tying ...
40. Why does context bloat occur when multiple MCP servers are connected simultaneously?
Each MCP server an agent connects to injects its own tool metadata, descriptions, and prompt templates into the model's context window, and those injections stack up as more servers are added. The model ends up sorting through far more tool options than are relevant to its current task Extra, unu...
41. What is the difference between sequential, hierarchical, and concurrent multi-agent orchestration?
Orchestration style How agents coordinate Sequential Agents run one after another, each building on the previous agent's output Hierarchical A coordinating orchestrator agent delegates subtasks to specialized worker agents Concurrent Multiple agents work on different parts of a task at the same t...
42. How can prompt injection bypass human-in-the-loop guardrails in an agentic workflow?
A human-in-the-loop checkpoint is designed to review the agent's proposed output before anything consequential happens, but some injection attacks are specifically crafted to trigger execution before that review ever occurs. Malicious instructions hidden in a document or data source the agent rea...
43. Why do enterprises prefer hardcoded state-transition graphs over freeform agent loops?
A freeform agent loop lets the model decide its own next step at every turn, which is flexible but makes the space of possible behaviors hard to predict or audit. A hardcoded graph defines the exact set of nodes and possible transitions in advance, so the model can only choose among pre-approved ...
44. When would you choose an SDK/microservice integration over MCP for tool access?
MCP's protocol-driven orchestration is most valuable when dynamic tool discovery and flexible, standardized coordination across many different tools and agents is a priority. If the main goal is simplicity and integrating with existing systems you already control, exposing agents as plain SDKs or...
45. Explain the execution flow of a ReAct agent handling a failed tool call?
sequenceDiagram participant A as Agent participant T as Tool A->>T: Tool call based on Thought T-->>A: Error or unexpected result A->>A: Observes failure, reasons about alternative A->>T: Retries with adjusted approach T-->>A: Success or further failure A->>A: Continues loop or escalates The agen...
46. Explain the internal working of a Plan-and-Execute agent's planning phase?
flowchart LR A[Goal Received] --> B[Task Decomposition] B --> C[Ordered Step Sequence Generated] C --> D[Steps Executed Sequentially] D --> E{Blocked?} E -->|Yes| B E -->|No| F[Task Complete] The agent receives the overall goal and, before taking any action, generates a full decomposition of that...
47. Explain the lifecycle of an agent's memory from short-term context to long-term storage?
flowchart LR A[Active Session: Short-Term Memory] --> B[Relevance Filtering] B --> C[Embedding or Structured Note Generation] C --> D[Stored in Vector DB: Long-Term Memory] D --> E[Retrieved via Similarity Search in Future Sessions] Active session : information starts in short-term memory, living...
48. Explain the internal working of Tree-of-Thoughts reasoning compared to Chain-of-Thought?
flowchart TD P[Problem] --> C1[Chain-of-Thought: single linear path] C1 --> A1[Answer] P --> T1[Tree-of-Thoughts: branch into candidates] T1 --> T2[Evaluate branches] T2 --> T3[Prune weak branches] T3 --> T4[Expand promising branches] T4 --> A2[Answer] Chain-of-Thought commits to a single linear ...
49. How can you optimize a multi-agent system to reduce token cost and latency?
Favor Plan-and-Execute over ReAct for sub-tasks with predictable steps, avoiding repeated reasoning cycles Use selective tool injection so each agent only sees the tools relevant to its specific role, reducing context bloat Cache and reuse retrieval results across agents working on the same task,...
50. How do you troubleshoot an agent that repeatedly calls the wrong tool?
Review the tool's schema and description, since the agent only ever sees the declared name and description, not the implementation, a vague or overlapping description is a common root cause Check for context bloat, if many tools with similar-sounding descriptions are all injected into context at ...
