📅 2025-04-18 — Session: Explored Modular AI Workflows with PromptFlow and RAG

🕒 18:40–19:20
🏷️ Labels: Ai Workflows, Promptflow, RAG, Architecture, Development
📂 Project: Dev
⭐ Priority: MEDIUM

Session Goal

The session aimed to explore the modular AI workflows using Microsoft’s PromptFlow and OpenAI’s Retrieval-Augmented Generation (RAG).

Key Activities

  • Discussed the modular approach to AI workflows, focusing on Microsoft’s PromptFlow and OpenAI’s RAG, detailing the roles of various tools and blocks in these frameworks.
  • Explored distinctions between Microsoft’s PromptFlow and OpenAI’s approaches, with a focus on embedding and retrieval tasks.
  • Outlined tools available in Microsoft’s PromptFlow for embedding generation, vector storage, and retrieval operations, including integration with RAG pipelines.
  • Architected an experience layer on PromptFlow to enhance workflows by integrating existing tools and adding custom abstractions.
  • Provided a detailed architecture map of the PromptFlow system, highlighting folder organization related to embedding, vector search, and reusable tools.
  • Outlined architecture map for FlowPower/AI Lambda Layer paradigm, detailing core block types, orchestration layers, and observability enhancements.
  • Explored the architecture and components of a Composable Semantic Runtime, detailing its modular AI workflow engine and runtime environment.
  • Completed the README for the FlowPower system, serving as a foundational introduction.

Achievements

  • Gained insights into the modular design of AI workflows and the integration of embedding and retrieval tasks.
  • Developed a comprehensive understanding of PromptFlow tools and their applications.
  • Completed foundational documentation for the FlowPower system.

Pending Tasks

  • Further develop documentation, templates, or starter code for the FlowPower system.
  • Explore actionable next steps for tool package creation in PromptFlow.