Blog
Practical writing on document parsing, RAG pre-processing, and Markdown for LLMs.
A PDF that looks perfect to a human often falls apart the moment a model reads it. Here's what goes wrong, and why converting to clean Markdown first makes downstream AI far more reliable.
Retrieval quality is decided long before the vector database. This is a pre-processing checklist for turning messy source documents into clean, chunk-ready text.
Plain text throws away the structure a model could have used. Markdown keeps it, at almost no token cost — and that changes how reliably an LLM can reason over your content.