RAG Engineer

Explain FActScore's approach to long-form factual evaluation and its design decisions

What does FActScore measure and how does it decompose long-form factual evaluation?

How is a RAG vector store used as an external long term memory layer inside an agent?

How does an agent access a RAG vector store without modifying its own weights?

How does agentic RAG differ from single shot RAG, and what does putting retrieval inside the loop enable?

What does agentic RAG uniquely enable compared to naive single shot RAG?

HNSW vs IVF, when do you pick each for a production vector index?

Which metric best measures whether a RAG answer is grounded in the retrieved context?

Why is cosine similarity preferred over Euclidean distance for text embeddings?

What is the primary purpose of embeddings in modern NLP?

Match each vector index type to its primary characteristic.

Put the steps of a basic RAG query pipeline in order.

What does RAG primarily help with in LLM-based applications?