ASAnvay Singhinanvay.hashnode.dev·Sep 5 · 11 min readI built a job-application copilot that can't lie about meI built Zapply, a personal job-application copilot. It finds jobs, ranks them against my resume, and drafts a tailored application. I paste and submit it myself. The design rule I kept everywhere: ev00
AAbhijatinexpensivetobewrong.hashnode.dev·Aug 31 · 12 min readThe Query Expansion That Made Retrieval WorseHere is a technique that will make your retrieval worse while looking, to everyone on the team, like it should have made it better. Take the user's query. Before you embed it, append a handful of syno00
AAbhijatinexpensivetobewrong.hashnode.dev·Aug 29 · 12 min readWe Doubled Retrieval Recall Without Touching the ModelThe most expensive retrieval bug I have shipped was not in the retriever. It was in the parser that fed it, and it was invisible to every metric I had. Here is the result, stated plainly and stripped 10
PNPradipta Nandainblog.zeonai.com·Aug 25 · 7 min readLate Chunking + Jina v4: Context-Aware Embeddings, Self-Hosted on GCPAuthors: Pradipta Nanda Amlan Das How we get retrieval that doesn't forget context, search that spans text and images, and full control over our embedding stack. If you've built a RAG system, you've73TIK
MSManu Shuklainecorpit.hashnode.dev·Aug 25 · 16 min readSentence Transformers 6.0 shipped ColBERT retrieval on 18 August 2026 and a 42x indexSentence Transformers 6.0 shipped ColBERT retrieval on 18 August 2026 and a 42x index Summary. Sentence Transformers v6.0.0 was published on 18 August 2026 at 13:46 UTC, twelve days after v5.7.0 on 6 00
IDInternals Decodedininternals-decoded.hashnode.dev·Aug 23 · 14 min readEmbeddings: Meaning as NumbersWhen you ask a chatbot for a pasta recipe, it doesn’t understand “pasta” the way you do. It sees a list of numbers. Those numbers are not random. They are arranged so that “pasta” sits close to “spagh00
RRajshreeinrjshree.hashnode.dev·Aug 22 · 25 min readYour RAG Isn't Broken. Your Retrieval Pipeline Is.author: "RAJश्री" A practical guide to diagnosing and improving Retrieval-Augmented Generation systems Most RAG systems don't fail because the LLM can't answer the question. They fail because the LLM00
CDCoding Dropletsincodingdroplets.com·Aug 16 · 11 min readSemantic Caching for LLM Calls in ASP.NET Core: When to Use It and HowEvery AI feature I have shipped eventually hits the same wall. Users ask the same twenty questions in a hundred different phrasings, and each one costs a full round trip to the model. Semantic caching24JK
DADaniel Anthony Romitelli Jrincraftedbydaniel.hashnode.dev·Aug 3 · 11 min readFour Vectors, One Record: How I Split Embeddings Before They Hit SearchThe failure that pushed me into this design wasn't subtle. My blended embedding kept returning candidates who looked fine at a glance and wrong on inspection. Ask about a very specific certification a00
DADaniel Anthony Romitelli Jrincraftedbydaniel.hashnode.dev·Aug 3 · 9 min readVector Split by Chunk: Why My Retrieval Stops at the Boundary I DrewA draft of mine missed the exact file span it needed. The vector was "close." But the chunk I wanted was buried inside a larger blob, and the miss was clean enough to be embarrassing: the sort of thin00
