Silent Failures: Why Production RAG Systems Degrade Without Throwing ErrorsA production RAG system can pass every operational check and still fail the only test that matters: telling the truth. Requests complete, latency stays within target, and the generated prose reads con17 min read·ahmetozel.hashnode.dev1Aug 21
Evaluating RAG: From Golden Datasets to LLM-as-JudgeA Retrieval-Augmented Generation system can produce a fluent answer while failing in several independent ways. It may retrieve the wrong evidence, omit a required document, misread correct context, an15 min read·ahmetozel.hashnode.dev1Aug 10
Beyond Vector Search: Knowledge Graphs, Structured Retrieval, and Intelligent RoutingVector search is an effective way to retrieve passages that are semantically similar to a question. It is not a universal interface to every kind of knowledge. Exact identifiers, deterministic calcula30 min read·ahmetozel.hashnode.dev2Aug 5
The Anatomy of Production-Grade RetrievalRetrieval-Augmented Generation is often summarized as a three-step workflow: split documents, embed the chunks, and send the nearest results to a language model. That description is useful for a proto34 min read·ahmetozel.hashnode.dev2Jul 21
Building a Local-First OCR + LLM Pipeline for Structured Business DocumentsHow to turn scanned invoices and contracts into validated data locally: OCR routing, schema-constrained extraction, and field-level validation.40 min read·ahmetozel.hashnode.dev24MJul 16
