LLM Agent Development using Semantic Kernel

Use Semantic Kernel in C# to develop secure, scalable AI agents with testing and monitoring capabilities.

Description

In this training, you will learn to use Semantic Kernel to leverage professional LLM Agents. Using a lot of hands-on exercises you will get familiar with the full scope of developing LLM Agents, including topics like LLMOps, prompt templates, API integrations, Retrieval Augmented Generation and memory management.

By the end of this training, you will be able to architect and develop secure, scalable AI agent systems using Semantic Kernel, with comprehensive testing and monitoring capabilities for business applications.

Learning Goals

• Understand large language models, their capabilities, and how to select the right model for different use cases [Understand]
• Construct and configure Semantic Kernel applications with multiple AI connectors in both console and web environments [Apply]
• Write effective prompts using templates, hyperparameters, and few-shot learning techniques [Apply]
• Explain the need for LLMOps practices including testing, monitoring, cost management, and security measures [Understand]
• Produce chat-based applications using conversation history and streaming responses [Apply]
• Implement testing and monitoring for LLM applications using OpenTelemetry and standard unit-testing tools. [Apply]
• Construct custom tools and functions that extend LLM capabilities, including external API integrations and filters [Apply]
• Implement Retrieval Augmented Generation (RAG) systems with vector stores for knowledge-enhanced AI applications [Apply]
• Employ structured output generation using JSON formatting and sideband communication patterns [Apply]
• Produce production-ready AI agents with proper tool integration, memory management, and security constraints [Apply]

Subjects

1. Introduction to Large Language Models
2. Configuring Semantic Kernel Applications
3. Effective prompting strategies
4. LLMOps
5. Building a chat-based application
6. Implementing testing and monitoring using OpenTelemetry
7. Extending LLM capabilities
8. Retrieval Augmented Generation (RAG) systems and vector stores
9. Structured output generation
10. Production-ready AI agents