Category: Key Terms
What is a semantic layer?
A semantic layer defines metrics and terms once for many BI tools. It is not a second warehouse. Start with one certified revenue metric before you buy a…
Fine-tuning vs RAG
Fine-tuning changes model behavior with examples. RAG retrieves documents into the prompt. Prefer RAG first for company knowledge; fine-tune when style or task format is the real gap.
What is an embedding?
Embeddings turn text into number lists so search and RAG can match meaning. Use them for similarity and retrieval. Do not treat a near neighbor as a verified…
What is a feature store?
A feature store shares ML feature definitions for train and serve so numbers stay consistent. Use one when skew hurts. Skip the platform when a spreadsheet of features…
What is grain in data?
Grain means what one row represents. Write it down before you join or sum. Mismatched grain is why three honest queries produce three revenues.
