The latest articles on DEV Community by Paz Gruberg (@paz_gruberg_3c403c00779df). https://dev.to/paz_gruberg_3c403c00779df https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F1851176%2F61722241-d1a8-43a4-8d94-efefdb19262b.png https://dev.to/paz_gruberg_3c403c00779df en Paz Gruberg Thu, 20 Feb 2025 12:55:05 +0000 https://dev.to/paz_gruberg_3c403c00779df/mvp-hack-loveabledev-ionic-capacitor-3ebg https://dev.to/paz_gruberg_3c403c00779df/mvp-hack-loveabledev-ionic-capacitor-3ebg <p>Hi all,<br> My co-founder and I are non-technical entrepreneurs with a strong background in product development. For our MVP, we're considering a “patchwork” approach by leveraging <strong>loveable (etc.)</strong> and then packaging it with <strong>Capacitor **and **Ionic</strong>.</p> <p>Since we're not developers, we realize this setup might seem like a quick hack rather than a robust, scalable solution - but scalability isn’t our focus right now.</p> <p><strong>Has anyone tried a similar approach or have tips on potential pitfalls and best practices? Any insights or alternative suggestions would be greatly appreciated.</strong></p> <p>Thanks in advance for your help!</p> ionic nocode javascript learning Paz Gruberg Sun, 04 Aug 2024 11:45:47 +0000 https://dev.to/paz_gruberg_3c403c00779df/seeking-advice-on-complex-ai-system-architecture-nlp-claude-parallel-flows-3h6b https://dev.to/paz_gruberg_3c403c00779df/seeking-advice-on-complex-ai-system-architecture-nlp-claude-parallel-flows-3h6b <p>I'm developing a system and could use some expert insights on its architecture and implementation. </p> <h2> <strong>Here's a brief overview of the current flow:</strong> </h2> <ol> <li>User fills out an intake questionnaire with X questions</li> <li>System references datasets/examples from existing system folders</li> <li>Multiple parallel flows run with Claude API to generate texts based on guidelines and relevant intake responses</li> </ol> <h2> <strong>Now, I'm grappling with some key issues:</strong> </h2> <ol> <li>Overall Architecture</li> <li>Should I use Agentic AI frameworks? If so, which ones?</li> <li>Or is there a more efficient approach for a system like this?</li> <li>Data Storage and Management</li> <li>How to store intake responses for efficient AI flow access?</li> <li>Is RAG (Retrieval-Augmented Generation) advisable for storage and retrieval? If not, what are the alternatives?</li> <li>Existing Dataset Integration</li> <li>What's the most efficient way to incorporate examples from existing folders into the flow?</li> <li>Better to train a small model on the dataset or integrate it directly?</li> <li>Parallel Process Optimization</li> <li>How to efficiently manage and synchronize the parallel flows?</li> </ol> <p><strong>Important notes:</strong></p> <ul> <li>Output (generated texts) and number of intake questions remain constant</li> <li>It's a complex flow split into several sub-flows, not a single long prompt</li> </ul> llm ai webapp learning