core practice · implementation task
Build a structure-aware chunker
Write a chunker that splits on document structure (headings/paragraphs) with a target size and overlap, instead of fixed character cuts.
The mechanism
A chunk is the atomic unit of retrieval — too big and embeddings blur, too small and context is lost. Good chunking respects structure (headings, functions, paragraphs) and adds overlap so answers are not cut in half.
Expected evidence
Chunks that respect structure and overlap at boundaries.
Try this drill in the practice workspace →
Hints when you need them
- Overlap prevents losing boundary-spanning answers.
- Different document types may need different strategies.
Watch for these mistakes
- Fixed character splits that cut sentences mid-thought
- Zero overlap, so boundary-spanning answers are lost
- One chunk strategy for every document type