Over the past weeks, I built and launched a sophisticated multi-agent retail operations platform powered by Google's Agent Development Kit (ADK) and Elastic's hybrid search—combining vector, semantic, and traditional search in real time. 👥 What makes it unique? - Multi-agent orchestration: Each specialized agent (Product Search, Review Analysis, Inventory, Shopping, Customer Support) works in harmony, managed by a central Coordinator Agent. - Conversational Intelligence: Handles natural language queries like “Show me blue summer dresses under $50 with good reviews in stock at my local store”—and delivers accurate, actionable results. - Cutting-edge tech: Google ADK, Gemini 2.0 Flash, Elastic kNN vector search, image-based product search (ImageBind embeddings), real-time dashboard UI, FastAPI backend, and more. - Real-world impact: Streamlines retail experience—instant product discovery, inventory insights, review analysis, FAQ support, and analytics. 🔗 Check out the codebase on GitHub:github.com/Yash-Kavaiya/AI-Accelerate-Retail-Agent-Teams 🔎 Full project details, technical deep dive & demo:devpost.com/software/ai-retail-agent-team-intelligent-retail-operations-platform Thanks to cutting-edge AI, this project brings together semantic/visual search, natural language capabilities, and enterprise-grade scalability for the next generation of retail operations! Would love to connect with retail tech enthusiasts, AI practitioners, and anyone interested in multi-agent systems for real-world applications. #ConversationalAI #GoogleADK #ElasticSearch #RetailTech #AI #MultiAgentSystems #GenAI #Gemini

As some of you will know I was creating an application only using prompting. Well ist all done now, so I thought i had better share it with you all. Remember im not a coder, UI/UX expert. I just used teh tools in the AI tool cheast (No I dont have a tool cheast fiull of AI tools, just Cursor, and the wider internet for ideas) Let me know what you think. Its not a polished video like Brandons, its 8 minutes of your life yyou may never get back. Code is freely available on Github - https://github.com/twelsh37/twospace User Demo at - https://assetms.theaiaa.com https://youtu.be/zUxr0iEIVYM

I have been 'Promptgramming' hard over the past few days /week to bring my Asset Management system to life. The drive behind this is my current clients shockingly bad IT Asset Management system. The clients Asset Management System is called onespace, and is a Microsoft PowerApps program. The Head of Operations has been writting this program for several years and changes its functionality on a daily, sometimes multiple times per day, basis. There is no dev system. All updates go straight to production. Testing? Us users do that and provide feedback via teams channel stating something has stopped working. Version control? Well i see version numbers change but im not sure how powerapps handles it. TFS, is that still a product? So i wrote my own as an experiment in the 'Art of the possible' My IDE runs on my Windows 11 PC, but because I am a geek/true professional I also use my Mac for development. I must note that I seem to get higher quality, better looking prompt feedback on the Mac compared to what I recieve on the PC. Why, I have no idea, just an observation. I use Cursor as my IDE, I have the paid for version. My Models are set to Auto and the Chat interface is set to Auto. I use git and Github for Version control - This enables me to easily pull the code between my two computers I am using Jest as my testing framework. - You could spend days on this. I spent all Saturday trying to get better coverage. Cursor got hung up on some problems and couldnt fix some errors. I could have persevered but in the end, i rolled the code back to the working state I had first thing Saturday morning. What a waste of a day that wa I just asked Cursor for an update about where we are with the program and got the following reply.... ________________________________________________ I'll analyze the project structure and current state to give you a comprehensive summary. Read file: masterplan.md Read file: PHASE2_README.md Read file: frontend/package.json