Back to AI Agents / RAG
About this interview
A technical interview on Agentic Design Patterns, pitched at the medium level. A voice AI interviewer leads the conversation, adapts its questions to your answers, keeps you on topic, and afterward gives you honest, specific feedback on where you were strong and where to improve. Expect roughly 30 minutes.
What you'll be assessed on
Explain why complex goals must be decomposed into ordered subtasks before execution
Describe structured output (JSON schemas) as the mechanism for machine-readable planning
Compare static plans versus event-driven replanning when unexpected inputs arise
Evaluate subtask outcomes and know when to iterate versus escalate
Topics covered
Task decomposition motivationTask decomposition basicsDecomposition benefitsPlanner-executor separationGoal clarityStructured output for planningStructured output schema designSchema-constrained outputStatic plansEvent-driven replanningStatic vs iterative planningSubtask outcome evaluationOutcome evaluation criteriaEvaluator pattern
A few sample questions
Just examples to set expectations - the real interview has many more and adapts to your responses.
“Why can't most real-world agent tasks be solved in a single LLM call, and what does that force us to do architecturally?
“After a subtask completes, how does the agent decide whether to move to the next subtask, retry the current one, or abandon the overall plan?
“Walk me through what happens step by step when a planning agent encounters a subtask it cannot complete because a required tool returns an error.