Will Velida - Microsoft

Building AI Agents with Dapr

Abstract

"Dapr Agents provides developers with a framework to build production-grade resilient AI Agents. Using Dapr Agents, we can build AI Agents that can work with Large Language Models to achieve complex tasks over stateful workflows.

In this talk, we'll talk about how we build effective AI Agents using Dapr by applying agentic patterns—ranging from augmented LLMs and prompt chaining to orchestrator-worker delegation and autonomous agents—using real-world examples and Python code to illustrate how these patterns can be applied to your applications.

Throughout the session, we’ll unpack how Dapr’s building blocks can simplify the orchestration and integration of tools, enabling developers to build scalable, observable, and resilient systems. We’ll also touch on deployment strategies and how Dapr’s abstraction layer enables switching between LLM providers without major refactoring.

By the end of the talk, you will not only understand how Dapr Agents work but also gain the confidence to apply these patterns in your own AI Agents using the Dapr Framework"

Speaker Bio

"Will is a Senior Software Engineer at Microsoft, working in the Industry Solutions Engineering team.

Will has extensive experience in software development, building solutions in the financial, health and agricultural industries. Will has also previously been a Microsoft MVP for Kubernetes and Open Source before joining Microsoft.

Will regularly speaks at conferences, user groups and developer events around the world. When he is away from his keyboard, you'll find him hiking around the countryside."