3 hints available

EasyAI EngineeringTheoryRAG Architecture Token Efficiency Vector Databases LLM Evaluation

Vector Database Architecture — Conceptual Questions

Answer each of the following questions about vector databases:

A startup is building a RAG system with 50,000 product descriptions. Their backend already runs on PostgreSQL. Their engineering team is 2 people. Should they use pgvector or a dedicated vector database like Pinecone? Justify your recommendation.
A user searches for "GPT-4 token limit" in a knowledge base. The correct document says "OpenAI GPT-4 supports a maximum context of 128,000 tokens." Another document discusses "billing limits for API token usage." Which document would pure semantic search likely return, and why might this be a problem? What does this suggest about retrieval strategy?
What does "approximate" mean in Approximate Nearest Neighbor search? What is being traded off, and why is this tradeoff acceptable in practice?

Sign in to attempt this problem

Free account gives you full access to community problems with the complete solution reveal: golden answer, senior walkthrough, and score breakdown, after submission.

Start free →Already have an account? Sign in