Shahzad Malik’s Post

How does #langchain compare with #llamaindex ? I implemented 4 toy applications with both #frameworks : 1. Connecting to a local LLM. 2. Indexing local files and building a RAG system. 3. Making a chatbot. 3. Converting it into an autonomous agent. Before building your own LLM-powered application, check out the differences between the two frameworks, so that you can avoid picking the wrong tool. I hope this post helps you make better decisions. #llm #chatbot #largelanguagemodels

learend about langchain and langchain ecosystem. also about agents and rag system.

Very interesting course that includes different examples of GenAI projects, Basics about LangChain and LLMs 🚀

"Chat with a document ?" or "Build your customized knowledge bases ?" Learn how to accelerate #Embedding and #LLM model with #OpenVINO for your #RAG system，just 3 steps： 1. Upload your text documents 2. Build the retriever/knowledge base 3. Chat with it https://lnkd.in/gpSgc_ZJ Powered by #LangChain

I was into innovative concepts around knowledge graphs, RAG, and LangChain. Inspired, I tried an idea for a Retrieval Augmented Generation on a Graph Database. I developed a demo showcasing a question-answering system utilizing Neo4j and LangChain to interact with a knowledge graph. Noteworthy, the entire demo was crafted using an open-source LLM. #Innovation #KnowledgeGraphs #Neo4j #LangChain #QuestionAnswering #OpenSource #LLM

Since I heard about LangChain and its capabilities to integrate any LLM (closed-sourced or open-sourced), I wanted to look into this framework. My curiosity about this technology has led me to write a how-to blog. It will work as a set of notes for me and others too. #ai #langchain #llms #softwaredevelopment

Completed a hands-on exercise on LangChain's "Chat With Your Data" project, where I dove deep into the potential of Retrieval Augmented Generation (RAG) and building responsive chatbots based on document content. 📊 Data Loading: Exploring over 80 data loaders to integrate various sources, including multimedia. 🔎 Data Splitting: Implementing best practices for optimal data handling. 🧠 Vector Stores & Embeddings: Understanding the core of embeddings and their integration for advanced data retrieval. 📚 Advanced Retrieval Techniques: Enhancing data indexing and retrieval to surpass traditional semantic queries. 🤖 Question Answering & Chatbot Development: Crafting a sophisticated one-pass QA system and maintaining conversation context in chatbot interactions. Notebooks: https://lnkd.in/gTSrATTi #LLM #RAG #Langchain

#DailyMoneta - I’ve explored different models, frameworks, and tools over the past week and a half. After extensive testing, I’m refining my approach to achieve the optimal balance between capability and simplicity. 🔍 What I’ve learned so far: The LangChain ecosystem is invaluable for streamlined AI workflows, but only some features align with my objectives. My focus has shifted from creating a generic toolchain to building specific tools that can make a difference. This refined approach will help me continue working on document creation with a broader, more scalable vision. I’ll roll out the initial workflows with my new insights in the coming days. Are you interested in learning more about my journey and the tools I’m using? I’m sharing detailed insights and practical takeaways on my blog. Head over to look at my process and stay updated on what’s next! 👉 https://lnkd.in/dvwweX4k #AIDevelopment , #LangChain , #WorkflowBuilding , #ToolchainDesign , #DocumentAutomation , #AIProject , #TechInsights

Is your Langchain debt serviceable? One quick litmus test to know if you are infact really building useful products with GenAI is the extend to which your codebase relies on packages like Langchain & Crew AI. If you are building extensively on top of them it is a red flag (imo) for 2 key reasons: 1. These frameworks are purposely designed to be restrictive (artificially) so that over time developers get locked in to specific design frameworks for their application and as a result become easier to upsell to (e.g. other pricier add-ons like langsmith). 2. The design-patterns for automating specific workflows can differ wildly based on usecase and it's very unlikely to be solved through generic & restrictive frameworks. Take a first principles approach! It's way too early to double-down on any of these packages. #genai #langchain #crewai #agents #work

💨 Meet Algebra, a DEX Engine with its latest forkable modular CLAMM codebase—Integral, akin to Uniswap V4, already trusted by major DEXes: #CamelotDEX, #THENA, #QuickSwap & beyond. 🎊 It's soon to be available on #SeiNetwork! 🔴 🧩 Algebra Integral is the first #V4 catalyst, featuring a #Modular Architecture that splits the monolithic #DEX setup into a core codebase & customizable plugins! 🔌 Plugins can include features like Dynamic Fees, On-chain Farming, Perps, AI & much more, allowing for quick implementation of almost any DeFi Functionality. 🧮 This approach opens up some opportunities for #DeFi builders, as well as facilitates rapid DEX updates, avoids liquidity migration, reduces gas costs by up to 22% on average & offers other benefits. 🪄 And Yaka Finance is going to be the first DEX deploying Integral on Sei Network, as the Sei's native liquidity engine! Stay tuned for more updates.