Beyond Prototypes: Building Production Ready AI Workflows

Building a quick AI prototype in .NET is easier than ever, but architecting an AI-driven backend that can handle production workloads is a different story. Long-running prompts, rate-limited external APIs, and complex state management quickly turn simple intelligent apps into distributed systems challenges. This session explores the idea of durable AI for modern systems. We'll look at building robust backends in C# that leverage libraries like Microsoft.Extensions.AI alongside Temporal's open source durable execution engine to orchestrate intelligent tasks reliably. We'll walk through code examples showing how to encapsulate LLM calls into resilient, stateful workflows that survive crashes, network issues, and API throttling, giving you a clear blueprint for production-ready AI.

Speakers

Cecil Phillip