In this hands-on workshop, we’ll build a real AI chat app and layer in enterprise-grade security using Google Cloud’s Model Armor — before shipping it to production on Cloud Run.

We’ll explore real-world attack vectors like prompt injection and system prompt hijacking, see how modern LLMs handle them, and then add Model Armor as a dedicated security layer that screens every prompt and response — independent of the model.

Finally, we’ll containerise the whole thing and deploy it to Cloud Run. You’ll walk away with a working, secured, production-ready AI app and a security pattern you can apply to any LLM project.

Plus there will be food & networking session at the end.