All interview questions
AI AgentsFoundations
What is an AI agent? How is it different from a regular LLM call?
Think first
Think about: tools, loops, autonomy, stopping conditions.
Strong answer
An AI agent is an LLM with tools, memory and a loop. A regular LLM call is single-shot: prompt in, response out. An agent receives a task, reasons about which tool to call, executes the tool, observes the result, and loops until it can produce a final answer or hits a stopping condition (max iterations, cost budget). The key differences: (1) Tools — the agent can call functions you define. (2) Loop — it iterates, not single-shot. (3) Autonomy — it decides what to do next. (4) State — it maintains context across steps. Production agents need guardrails: max_iter, cost budget, per-tool timeout, tool-call dedup, tracing, and human-in-the-loop for destructive actions.
Weak-answer indicators
- Says 'an agent is a smarter LLM'
- Doesn't mention the loop or tools
- Skips stopping conditions / guardrails
- Confuses agents with RAG
Likely follow-up questions
- What is the ReAct pattern?
- How do you prevent an agent from looping forever?
- How do you secure an agent with tool access?