<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Megha Sharma on Medium]]></title>
        <description><![CDATA[Stories by Megha Sharma on Medium]]></description>
        <link>https://medium.com/@megha.sharma3333?source=rss-af736ae55d64------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*rAVwVEl0Pl-xmACCi0TITw.jpeg</url>
            <title>Stories by Megha Sharma on Medium</title>
            <link>https://medium.com/@megha.sharma3333?source=rss-af736ae55d64------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 03:24:45 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@megha.sharma3333/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Retrieval Augmented Generation(RAG): A Beginner’s Guide]]></title>
            <link>https://blog.gopenai.com/retrieval-augmented-generation-rag-a-beginners-guide-13d1e7102e2f?source=rss-af736ae55d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/13d1e7102e2f</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Megha Sharma]]></dc:creator>
            <pubDate>Thu, 17 Jul 2025 20:47:23 GMT</pubDate>
            <atom:updated>2025-07-19T16:28:56.361Z</atom:updated>
            <content:encoded><![CDATA[<p>Imagine there is a student (The LLM) preparing for a debate competition who is brilliant, articulate and intelligent and well trained through his textbooks through the years but if he is asked about some recent event he might “hallucinate” i.e he might come up with some answer but that’s not accurate as he is not updated.</p><p>Now that’s where the RAG(Retrieval Augmented Generation) comes into play- it’s like equipping the student with a personal real-time library during the debate where he can answer to any question that is latest or complex which is not updated in his textbooks or the documents he went through earlier. The RAG will help him go through all the relevant articles, reports and documents from that library.</p><p>In simple words, RAG is a powerful AI framework that enhances the capabilities of LLMs(Large Language Models) by enabling them to utilise the external, up-to-date and domain specific beyond the models or data they were originally trained on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qpI9Rh5FhjJD--cJ95QLNA.png" /><figcaption>RAG Pipeline</figcaption></figure><h3>Working Of RAG</h3><p>Let’s break it down in four phases:</p><p><strong>Ingestion</strong></p><ul><li>This will be our first step where we take our data from external sources(documents, databases, API’s, websites, PDF’s, CSV’s etc).</li><li>Chunking: We need to break down our data or document into smaller chunks to have a limited context window. The smaller chunks ensure the relevant pieces of information that can be retrieved and fit within this window.</li><li>Vector Embeddings: These chunks now need to be converted into a numerical representation called “vector embeddings”. These vector embeddings have to be indexed and stored in vector databases (Pinecone, ChromaDB, Milvus, FAISS).</li></ul><p><strong>Retrieval</strong></p><p>Now this phase comes when the user submits a query. The user’s query is also converted into vector embeddings we do this to quickly search the vector databases. It retrieves the most relevant chunks from the knowledge base. There are different retrieval methods like Maximum Marginal Relevance, a simple cosine similarity or even a hybrid search.</p><p><strong>Augmentation</strong></p><p>The retrieved context chunk is then augmented with the user’s original user prompt forming an “augmented prompt”.</p><p><strong>Generation</strong></p><p>The augmented prompt is sent to the LLM which generates a grounded response.</p><p>So this was RAG in brief. Looking ahead there’s a lot more to explore from some powerful frameworks like LangChain to build robust AI Agents.</p><p>I hope you could find some value in this article. Thank you for your time and interest.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=13d1e7102e2f" width="1" height="1" alt=""><hr><p><a href="https://blog.gopenai.com/retrieval-augmented-generation-rag-a-beginners-guide-13d1e7102e2f">Retrieval Augmented Generation(RAG): A Beginner’s Guide</a> was originally published in <a href="https://blog.gopenai.com/">GoPenAI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>