Interview Q&A Quick Reference
ML/GenAI Knowledge Q&A covers how to prepare for the rapid-fire breadth round. This page is the actual drill: real interview questions, click to reveal a concise answer, then check yourself against it. Each answer links to the full page if you need the real depth behind it — the point here is testing recall under a few seconds of pressure, not re-teaching the concept from scratch, so keep answers short and move on.
Three categories to start (LLM fundamentals, RAG, and agents — the highest-traffic combination for AI engineering interviews right now).
LLM Fundamentals
❓ What is a foundation model, and how is it different from a regular ML model? ▶
❓ What is a Large Language Model, and what is it actually trained to do? ▶
❓ What is tokenization, and why can't LLMs just read raw text? ▶
❓ What is Byte Pair Encoding (BPE)? ▶
❓ What are embeddings? ▶
❓ What is positional encoding, and why do Transformers need it? ▶
❓ What is self-attention? ▶
❓ What are Query, Key, and Value in attention? ▶
❓ Why do we scale attention scores by the square root of the key dimension? ▶
❓ What is multi-head attention, and why not just use one big attention head? ▶
❓ What's the difference between self-attention and cross-attention? ▶
❓ What is masking used for in a Transformer? ▶
❓ What is a KV cache, and why does it matter for serving? ▶
Retrieval-Augmented Generation (RAG)
❓ What is RAG, and why not just fine-tune the model on your data instead? ▶
❓ What are the main stages of a RAG pipeline? ▶
❓ What is chunking, and why does chunk size matter? ▶
❓ What's the difference between dense and sparse retrieval? ▶
❓ What is BM25? ▶
❓ What is hybrid search, and why combine dense and sparse retrieval? ▶
❓ What is re-ranking, and why is a second retrieval stage needed at all? ▶
❓ What's the difference between a bi-encoder and a cross-encoder? ▶
❓ What is GraphRAG, and when does plain chunk-based RAG fail that it fixes? ▶
❓ How do you evaluate a RAG system? ▶
❓ Why might increasing top-k retrieved chunks hurt instead of help? ▶
AI Agents
❓ What is an AI agent, and how is it different from a single LLM call? ▶
❓ What is tool use / function calling? ▶
❓ What is the ReAct pattern? ▶
❓ What is Plan-and-Execute, and how does it differ from ReAct? ▶
❓ What is reflection / self-critique in an agent? ▶
❓ What's the difference between single-agent and multi-agent systems? ▶
❓ What are the three multi-agent coordination topologies? ▶
❓ What is the supervisor / routing pattern? ▶
❓ What is MCP (Model Context Protocol), and what problem does it actually solve? ▶
❓ What are the agent memory types? ▶
❓ What is context engineering, and how is it different from prompt engineering? ▶
❓ What is context rot? ▶
❓ What's a sub-agent architecture, and why give a sub-agent its own context window? ▶
Next: Technology Comparisons & Decision Trees — the "why X over Y" tradeoff questions this same rapid-fire round often includes.