In this workshop, participants will build a production-style “ChatGPT” application using Azure OpenAI and C#. Starting from a clean .NET solution, we will design and implement:

A chat orchestration layer

Conversation memory with token-aware trimming

A project-based knowledge engine that can ingest 20–40 local files

Embedding generation and vector similarity search (without Azure AI Search)

Prompt injection for grounded responses

A Web API layer suitable for real-world deployment

Production considerations including ingress, rate limiting, and cost control

Attendees will implement a miniature Retrieval-Augmented Generation (RAG) system using in-memory vector search, learning how modern AI assistants are constructed under the hood.

The session concludes with a comparison between local vector storage, SQL vector capabilities, and enterprise-scale Azure AI Search architecture — providing a clear upgrade path from conference demo to production system.

This workshop is intended for senior developers, cloud engineers, and solution architects who want to move beyond simaple AI API calls and understand the architectural patterns required to build secure, scalable AI applications in the enterprise.

Build a production-style “ChatGPT” application in C# using Azure OpenAI. In this hands-on workshop, participants will implement chat orchestration, conversation memory, and a project knowledge engine capable of ingesting 20–40 files using embeddings and vector similarity search — without Azure AI Search. The session concludes with deployment, security, and enterprise upgrade patterns for scaling to production.