Content Filtering Pipelines for Web-Retrieved LLM Context
Cleaning retrieved documents matters as much as the model itself.
Saoirse Ó Briain
Staff Writer
Saoirse Ó Briain
Saoirse came to technical writing through a linguistics PhD and a stint at a Dublin-based AI startup working on grounding and factuality in LLM outputs. She is the publication's lead voice on context engineering, covering how retrieval shapes model behavior and where the discipline is heading.
6 stories
Why SERP APIs Return Insufficient Context for LLM Reasoning
SERP APIs return snippets and links, not the full text models need to reason.
Saoirse Ó Briain
Metadata Filtering in Vector Databases for RAG
Filtering metadata before vector search prevents irrelevant results from reaching your model.
Saoirse Ó Briain
RAG Evaluation Metrics Beyond RAGAS for Web Knowledge Tasks
RAGAS misses four critical dimensions web retrieval demands.
Saoirse Ó Briain
Re-Ranking Web Retrieved Documents for LLM Context Windows
Reranking alone won't save your RAG system without strategic document placement.
Saoirse Ó Briain
Hybrid Retrieval Combining Dense and Sparse Indexes for Web Search
Combining keyword and semantic search beats either approach alone on diverse queries.
Saoirse Ó Briain