Data/ML/AI
Chunking
Also written as Document Chunking, Chunking Strategy
Splitting documents into smaller pieces before they are embedded and stored for RAG, so the system can fetch just the relevant passage instead of a whole manual. How the text is split has a big effect on answer quality.
Think of it like
Cutting a cookbook into individual recipe cards so you can grab the one you need without the whole book.
Junior or senior?
Junior sounds like
Used a library's default chunk size.
Senior sounds like
Tried different strategies (by heading, by paragraph, with overlap), measured the effect on retrieval, and can say what worked for their documents.
Ask them
“How did you decide how to split your documents, and how did you know the choice was a good one?”