LLM Fundamentals

Prompt Engineering

RAG

Vector Database

RAG vs Fine-Tuning vs Long context

RAG Question

Fine-Tuning

Quantization

Safety & Guardrails

Agentic System

What to drop in every multi-agent answer

A condensed checklist:

1. Justify multi-agent over single — specialization, isolation, parallelism, model tiering
2. Name the pattern — supervisor-worker, ReAct, hierarchical delegation
3. State management approach — typed schema, checkpointing, what's shared vs private
4. Cost story — model routing, context pruning, caching
5. Failure modes — loops, compounding error, hung agents — with mitigations
6. Observability — distributed tracing, per-agent attribution
7. Eval strategy — end-to-end success + per-agent + trajectory