Hey there!

Come join me as I start my prep for the Agentic AI Hackathon hosted on the 25th of August (just a few more days!!)

I hear the term Agents from colleagues but it hasn’t peaked my interest just yet because of the huge pile of list on my bucket. I didn’t want to add more to the list! 😅

Let’s now build a prep plan together with Copilot!

The recommended learning & hands-on Plan:

Phase 1: Foundations

Phase 2: Hands-On Build & Explore

Phase 3: Deep Dive

  • Goal: Build prototypes to gain more experience
  • Actions:
    • Azure AI Agents:
      • Build and deploy a single agent with real actions.
    • End-to-End Multi-Agent:
      • Build an agent chain that uses orchestration patterns.
    • Enable Multi-Agent Frameworks:
      • Focus on framework-level integration (SK + AutoGen).
    • Presentation Builder:
      • Pair agent orchestration with document generation.
  • Prep:
    • Sketch architecture: determine agents, tools, interactions.
    • Build minimal viable prototype with Semantic Kernel and/or AutoGen.
    • Use Azure Foundry SDK (C# or Python).
    • Add output layer such as web interface / presentation export.
    • Incorporate security and governance basics (human-in-loop, input sanitisation)

Let me know your thoughts and here’s to a fun and exciting hacking!