Technical interview guide

Troubleshooting LLM, RAG & Agent Systems

A practical diagnostic playbook for the specific ways LLM, RAG, and agent systems fail in production.

Read: 45 min
Practice MCQs: 25
Interview QA: 25
Edition: v4
Editorial status: Reviewed
Relevant for: AI EngineerAgentic AI EngineerForward Deployed Engineer (FDE)

Scope: RAG, DPR, Lost in the Middle, RAGAS, ReAct, Toolformer, MemGPT, AutoGen, MT-Bench, OpenTelemetry, W3C Trace Context, Model Cards, NIST GenAI, and OWASP Prompt Injection/Excessive Agency references reviewed 2026-09-04.

Interview QA

Treat each question like a live interview question: answer out loud first (structure, assumptions, tradeoffs), then open the model answer to spot gaps and rehearse a tighter follow-up.

Curated: 2026-10-05 · Written: 2026-10-05 · Reviewed: 2026-10-03

QA-1

Triage a severe LLM production incident.

QA-2

How do you diagnose and break infinite tool-calling loops or non-terminating planning cycles in autonomous agent systems in production?

QA-3

How do you debug multi-agent coordination failures such as semantic deadlocks, goal drift, and conflicting concurrent tool actions?

QA-4

Reduce a complex agent failure to a regression case.

QA-5

Troubleshoot a prompt-template regression.

QA-6

Troubleshoot a structured-output failure.

QA-7

Diagnose and fix generation truncation.

QA-8

Troubleshoot an underperforming RAG answer layer by layer.

QA-9

Diagnose a RAG no-result incident.

QA-10

Troubleshoot an embedding migration regression.

QA-11

Troubleshoot unsupported RAG citations.

QA-12

Diagnose an indirect prompt-injection event.

QA-13

Troubleshoot an agent loop.

QA-14

Diagnose an ambiguous agent side effect.

QA-15

Troubleshoot a tool authorization failure.

QA-16

Troubleshoot wrong personalization from memory.

QA-17

Diagnose high p99 latency in an LLM-agent service.

QA-18

Troubleshoot an LLM cost regression.

QA-19

Troubleshoot a sudden automated-evaluation score drop.

QA-20

Troubleshoot an observability blind spot.

QA-21

Diagnose apparent LLM behavior drift.

QA-22

How do you troubleshoot context window overflow and token starvation during complex multi-step reasoning runs?

QA-23

Write the closure criteria for an LLM-system incident.

QA-24

Design rollback for a failing LLM/RAG/agent release.

QA-25