SKSerhiy Kucherenkoinskucherenko.hashnode.dev·Aug 25 · 4 min readHow to hurt yourself with AISummary 34x more work than it should have been. Timeline I write technical articles, and I like experimenting with something new in each one. For one of them, the experiment was an AI pipeline: it wou00
SKSerhiy Kucherenkoinskucherenko.hashnode.dev·Aug 14 · 4 min readJudge strictly to avoid hallucinationsTen questions over the SEPA rulebooks. A second model grades each answer against a reference, zero to a hundred, and the run came back healthy. Except for question five, which asked which currency SC00
SKSerhiy Kucherenkoinskucherenko.hashnode.dev·Aug 9 · 4 min readThe illusion of improvementThe problem My RAG answers questions about SEPA payment rulebooks. Every page of those PDFs carries the same header and footer, and all of it lands inside the chunks that get embedded. Two lines of l00
SKSerhiy Kucherenkoinskucherenko.hashnode.dev·Aug 2 · 4 min readThree small lessons from building a RAG by handA retrieval system over the SEPA payment rulebooks. Plain Python, Postgres, two LLM vendors. Three independent decisions from it are worth writing down, because in each case the decision was cheap and00
SKSerhiy Kucherenkoinskucherenko.hashnode.dev·Jul 27 · 5 min readThe localhost trap: a 10-second database connection on WindowsEvery question in the app took about 15 seconds. The model was fast. Retrieval was fast. Ten of those seconds were hiding somewhere else, and for days I couldn't say where. The app is payments-rag, a 00
