---
title: "Harshal Chauhan (@harshal8140) | Hashnode"
description: "Harshal Chauhan (@harshal8140) writes on Hashnode. Read their latest articles and follow their work."
url: "https://hashnode.com/@harshal8140"
---

@harshal8140

# Harshal Chauhan

@harshal8140 · Joined July 2025

### [4.1 HLOV - Shrinking Ten Million Needles: Quantization, and Why Filters Break Vector Search](https://retrieval-augmented-generation-guide.hashnode.dev/4-1-hlov-shrinking-ten-million-needles-quantization-and-why-filters-break-vector-search?utm_source=hashnode&utm_medium=feed)

Priya's HNSW index is fast and accurate, but it has two new problems: Memory. 10 million chunks × 3,072 bytes is 30.7 GB of RAM, and the archive grows by a million chunks a month. The fix is quantiza

11 min read · [retrieval-augmented-generation-guide.hashnode.dev](https://retrieval-augmented-generation-guide.hashnode.dev/4-1-hlov-shrinking-ten-million-needles-quantization-and-why-filters-break-vector-search?utm_source=hashnode&utm_medium=feed)

[2d ago](https://hashnode.com/posts/4-1-hlov-shrinking-ten-million-needles-quantization-and-why-filters-break-vector-search/6afc8d182d431d793ec85ef8)

### [3.1 HLOV - Searching Ten Million Needles in Ten Milliseconds: ANN Indexes](https://retrieval-augmented-generation-guide.hashnode.dev/3-1-hlov-searching-ten-million-needles-in-ten-milliseconds-ann-indexes?utm_source=hashnode&utm_medium=feed)

The search that got slow Priya's hybrid search is working well, so her VP decides to add a lot more data: six years of claim notes, emails, and call transcripts. Over one weekend, the index grows from

17 min read · [retrieval-augmented-generation-guide.hashnode.dev](https://retrieval-augmented-generation-guide.hashnode.dev/3-1-hlov-searching-ten-million-needles-in-ten-milliseconds-ann-indexes?utm_source=hashnode&utm_medium=feed)

[2d ago](https://hashnode.com/posts/3-1-hlov-searching-ten-million-needles-in-ten-milliseconds-ann-indexes/6afc746167cd6dbbb9133663)

### [2.1 HLOV - Finding the Needle: Sparse Retrieval, Dense Retrieval, and Hybrid Search](https://retrieval-augmented-generation-guide.hashnode.dev/2-1-hlov-finding-the-needle-sparse-retrieval-dense-retrieval-and-hybrid-search?utm_source=hashnode&utm_medium=feed)

Two bugs, one cause Priya's policy assistant has been live for three weeks. Two bug reports arrive on the same morning: An underwriter asked about the GP-7 rider and got an answer about the GP-2 ride

15 min read · [retrieval-augmented-generation-guide.hashnode.dev](https://retrieval-augmented-generation-guide.hashnode.dev/2-1-hlov-finding-the-needle-sparse-retrieval-dense-retrieval-and-hybrid-search?utm_source=hashnode&utm_medium=feed)

[2d ago](https://hashnode.com/posts/2-1-hlov-finding-the-needle-sparse-retrieval-dense-retrieval-and-hybrid-search/6afc5aec7615483157e28bca)

### [17\. Self-RAG and Corrective RAG: Teaching RAG to Check Its Own Work](https://retrieval-augmented-generation-guide.hashnode.dev/17-self-rag-and-corrective-rag-teaching-rag-to-check-its-own-work?utm_source=hashnode&utm_medium=feed)

Basic RAG has a simple flow: take the question, retrieve the top few documents, give them to the LLM, and get an answer. It works well in demos. In production, it fails in a quiet and dangerous way. T

10 min read · [retrieval-augmented-generation-guide.hashnode.dev](https://retrieval-augmented-generation-guide.hashnode.dev/17-self-rag-and-corrective-rag-teaching-rag-to-check-its-own-work?utm_source=hashnode&utm_medium=feed)

[2d ago](https://hashnode.com/posts/17-self-rag-and-corrective-rag-teaching-rag-to-check-its-own-work/6afc49f0a87f3bc68ebd1184)

### [16.RAPTOR: Building a Tree of Summaries So Retrieval Can Answer Big-Picture Questions](https://retrieval-augmented-generation-guide.hashnode.dev/16-raptor-building-a-tree-of-summaries-so-retrieval-can-answer-big-picture-questions?utm_source=hashnode&utm_medium=feed)

Normal RAG retrieves small pieces. Some questions need the whole picture. RAPTOR groups related chunks, summarises each group, then summarises the summaries, and searches all of it together. Learn how

17 min read · [retrieval-augmented-generation-guide.hashnode.dev](https://retrieval-augmented-generation-guide.hashnode.dev/16-raptor-building-a-tree-of-summaries-so-retrieval-can-answer-big-picture-questions?utm_source=hashnode&utm_medium=feed)

[3d ago](https://hashnode.com/posts/16-raptor-building-a-tree-of-summaries-so-retrieval-can-answer-big-picture-questions/6afb3db4aa4ddaa1780598a2)
