“Explain embedding vectors and cosine similarity to me using a spatial analogy — no maths notation, just intuition. I want to be able to explain this concept to a non-technical colleague tomorrow.”
“I’m choosing between OpenAI’s text-embedding-3-small and text-embedding-3-large for a RAG system over legal documents. What factors should I consider, and is the higher cost of large likely to be worth it for this use case?”
“What is HyDE (Hypothetical Document Embedding) and when should I use it? Show me a Python implementation using the OpenAI API that generates a hypothetical answer to a query and uses that for retrieval instead of the raw query.”
“My RAG system is retrieving irrelevant chunks for some queries even though the correct information is definitely in my document collection. Walk me through how I would diagnose whether the problem is in the embedding model, the chunk size, or the similarity threshold.”
“I need to embed 50,000 document chunks using the OpenAI Embeddings API. Show me a production-grade Python implementation with batching, rate limit handling, retry logic, progress tracking, and cost estimation before the job runs.”
