Mike Brummitt’s Post

Looking to build a multi-query RAG pipeline? See how it can be done in 20 minutes using #Langflow. ⬇️ #RAGApplications #AIApplications #Langflow https://ow.ly/hyeg50TlCew

Looking to build a multi-query RAG pipeline? See how it can be done in 20 minutes using #Langflow. ⬇️ #RAGApplications #AIApplications #Langflow https://ow.ly/65fx50Tmc70

Looking to build a multi-query RAG pipeline? See how it can be done in 20 minutes using #Langflow. ⬇️ #RAGApplications #AIApplications #Langflow https://ow.ly/mBI650TsSym

Looking to build a multi-query RAG pipeline? See how it can be done in 20 minutes using #Langflow. ⬇️ #RAGApplications #AIApplications #Langflow https://ow.ly/I0jM50TuaEL

Looking to build a multi-query RAG pipeline? See how it can be done in 20 minutes using #Langflow. ⬇️ #RAGApplications #AIApplications #Langflow https://ow.ly/IhOM50Tm6NZ

Open-source multi-Agent platform? Mahyar Osanlouy, PhD.. thanks for the share! Need I say more? .. nope. Edit: Well actually.. I will. The whole git repo is full of awesome! ☺️ For example; FaceChain is a deep-learning toolchain for generating your Digital-Twin. 🔴 warnings..? Thanks to an anonymous source; There does appear to be a problem with the repo. Far too many stars for the amount of issues. Along with other indicators suggesting this could be problematic and needs solid vetting. 🟢 if all is good.. As Mike Hall notes in the comments, there are others out there; autogen studio & crewai being the popular ones. I wonder how it stakes up. #ai #aistrategy #llm #aichallenges #opensource #ai #mlops #ml #genAI #aritificialintelligence #aiagents

Excited to introduce ReasonFlow, the next evolution in structured AI workflows! At ReasonWorks AI, we’ve built ReasonChain, a powerful framework leveraging Chain-of-Thought (COT) and Tree-of-Thought (TOT) reasoning to solve complex, logical problems with precision. While ReasonChain is ideal for structured decision-making and planning, we realized there’s a growing need for flexible, low-level, and highly customizable workflows in dynamic AI environments. That’s where ReasonFlow steps in. What’s the difference? - ReasonChain focuses on structured, logical problem-solving with step-by-step reasoning. It’s perfect for industries requiring deterministic solutions and predefined workflows. - ReasonFlow, on the other hand, is all about customizable agentic workflows. It’s a low-level framework designed to integrate ReasonChain and RAG (Retrieval-Augmented Generation) seamlessly, enabling dynamic, end-to-end workflows that adapt to real-world complexity. Why ReasonFlow? With ReasonFlow, we aim to bridge the gap between rigid workflows and flexible AI pipelines. It allows businesses to create bespoke AI solutions that are modular, scalable, and adaptable for diverse use cases—from automation to advanced decision-making systems. Coming Soon: ReasonFlow Studio! I’m thrilled to announce that we’ll soon be launching ReasonFlow Studio, a drag-and-drop interface where users can: - Generate Agents effortlessly. - Design and run workflows seamlessly. - Visualize and optimize AI processes in real-time. This no-code solution will make it easier than ever to harness the power of ReasonFlow for creating advanced AI workflows. Open Source and Ready to Explore! ReasonFlow is now available on PyPI for easy installation and use. We’ve also made the project open source to foster collaboration and innovation. 📦 Install ReasonFlow: pip install reasonflow 🔗 Check out the GitHub Repository: https://lnkd.in/grgBN-eX This is just the beginning! Stay tuned for updates on ReasonFlow Studio and more innovations from ReasonWorks AI. 👀 Let me know your thoughts or suggestions—we’d love to hear your feedback! 🙌 #AI #MachineLearning #ReasonFlow #ReasonChain #AIWorkflow #ArtificialIntelligence #TechInnovation #NoCode #OpenSource #PyPI #AIFramework #RAG #DragAndDrop #AIStudio #ReasonWorksAI #Innovation #TechSolutions #WorkflowAutomation #AIIntegration #AITools #TechCommunity

BUILDING RAG FROM SCRATCH (OPEN SOURCE) The wonders of AI, like ChatGPT, have shown us the incredible potential of training computers on vast amounts of information. We're fortunate to witness this exponential technological shift in our lifetime. As many of you know, Large Language Models (LLMs) can process vast amounts of data and provide accurate answers to our questions. However, these models often lack one crucial element - YOUR DATA. Imagine harnessing the power of LLMs but using your documents, PDFs, databases and websites. You can ask a question to a 100-page document and receive the most relevant and accurate response. This is the promise of Retrieval-Augmented Generation (RAG). RAG is invaluable as it saves time, reduces costs and ensures we remain well-informed. Inspired by #llamaindex, I embarked on building a RAG solution using only open-source software and hosting parts of it on Google Cloud's free tier. The process was relatively straightforward, though I encountered some challenges with basic retrieval and chunking techniques. This demonstration of the technology serves as an ideal starting point, with more improvements and examples to come in the following days. If you need assistance with your RAG implementation, feel free to reach out. Check out the detailed solution in the video attached: [YouTube Video](https://lnkd.in/eYsxCcVT). Tech Stack: - LlamaIndex - Llama 2 - GCP Compute Engine - VSCode - Remote-ssh Extension - Huggingface 🤗  - Embedding model: BAAI/bge-small-en #AI #RAG #OpenSource #RealEstate #Technology #Innovation #Productivity #InformationRetrieval

RAG Developers, take note 🔍 This exercise showed that larger models can outperform smaller ones in precision and knowledge breadth. Choosing the right embedding can truly make a difference. Read more from Nathan Bos now. #LLM #Embedding #SemanticSearch #RAG

Just came across this company Ragie, interesting to me that a methodology such as RAG is being commodified. Also another example of Typescript being used in the space of #GenAI. I’ve been seeing Typescript pop up in more and more places as it relates to GenAI. I wonder if this is a larger push to have devs who typically work in the realm of FE development with languages such as #Typescript becoming the owners of the end-to-end GenAI dev cycle https://www.ragie.ai