Day 36

What I learned about Chunking strategies.

Documents need to be split into smaller chunks before they can be used in RAG.

I learned about two common approaches:

1. Fixed-size chunking
Split text by a predefined limit of characters, words, tokens, or sentences.
→ Simple and efficient, but it can break an idea in the middle.

2. Semantic chunking
Split text when the meaning or topic changes.
→ Better preserves context, but requires more processing.

My takeaway: How we split a document matters because it can affect the context RAG gets when retrieving information.

Day 36 of 1% better.