LLM App Development: Semantic Kernel Workshop

Learn how to build LLM Apps with Semantic Kernel through self-contained and independent lessons, in a flexible way so that you can choose the sequence that is most convenient for you.

Typically, each lesson is structured to last approximately one hour, and it is assumed that you already have some Python or coding skills for the hands-on labs.

Objectives

After completing this workshop, you will be able to:

Author your own copilot using Semantic Kernel.

Create AI plugins using semantic and native functions.

Automate complex tasks execution with planners.

Create prompt templates to define AI functions.

Embed Generative AI in your applications.

1 Background - Introduction to LLMs

no-code. 1.5h duration

Introduction to LLMs GPTs and other models.
Azure AI Services Overview.
Azure OpenAI Service Overview.

2 CoPilot Stack and Semantic Kernel

full-code. 1.5h duration

Introduction to Copilots.
Microsoft Copilots.
Build your own Copilots.
Copilot ecosystem (Copilots + Plugins).
What is Semantic Kernel?
Why Semantic Kernel?
The role of the kernel.

3 Semantic Kernel Basic Concepts

full-code. 1.5h duration

Native and Semantic Functions.
Create AI Plugins from functions.
Chain plugins together.

4 Semantic Kernel Advanced Topics

full-code. 3h duration

Add Memory to your AI Apps.
Use Planners to automate Plugins orchestration.
Calling external connectors.

5 Best Practices and Lessons Learned

no-code. 0.5h duration

Learn some best practices on service limits.
Final discussions and wrap-up.

Requirements

Workstation

Azure CLI
Anaconda
VS Code

Cloud

Azure Subscription

Download Details:

Author: microsoft

Official Github: https://github.com/microsoft/sk-workshop

License: CC-BY-SA-4.0, MIT licenses found

#microsoft