⛵ The meetup with a view - Cape Town Developer User Group.

Have you ever attended a meetup with the perfect view of the Atlantic Ocean? Speaking right next to where you board the ferry to the history-filled Robben Island? Yeah, I never thought I would—but I did. And afterward, I took a boat ride to Robben Island to immerse myself in its history.

Session: Building AI Solutions in Azure AI Foundry

For this meetup, I stuck to the philosophy: if it’s not broken, it doesn’t need fixing. So, I redelivered the same session I did in Joburg. The difference? More questions, more interaction, and an even more engaged audience. Plus, I got the chance to attend my colleagues’ sessions, which was a great learning experience.

This session serves as a perfect introduction to Azure AI Foundry and how to build AI solutions. For this blog, I want to focus on one key part of the session—Retrieval-Augmented Generation (RAG).

Retrieval Augmented Generation (RAG)

RAG is a technique that enhances LLM responses by incorporating your own data. Here’s how it works:

image.png

  • Data Sourcing and Formatting: Collect and store your data in one location.
    • Depending on its size, you may need to break it into smaller chunks—a process called chunking.
    • To make the data understandable for the model, it must be converted into a numerical format, known as vectorization.
    • Once vectorized, a search index makes the data easily searchable—just like a library catalog.
    • Finally, to ensure relevant results, the retrieved data is ranked based on importance.
  • Retrieval: When a user asks a question, the app searches the database for relevant information.
  • Augmentation: The retrieved data is added to the prompt to enhance the response.
  • Generation: The enhanced prompt is fed to the LLM, which generates a more accurate and context-aware response.

image.png

Sounds simpler now, right? You can learn more about implementing RAG in our Generative AI for Beginners course.

A session this good should be delivered again, right?

And that’s exactly what I did! In the next blog, I’ll share a live redelivery of the session so you can learn alongside me. See you there!

P.S. The next session is part of the Azure AI Fundamentals Series on Microsoft Reactor 📺