Google Cloud × NVIDIA Developer Meet-up & Reception at KubeCon

Your agent plans with a reasoning model, calls tools, pulls context from a retrieval pipeline, and maybe reads a screenshot or a PDF along the way. That's not one model behind one endpoint. It's a system of models, and each one has its own needs for GPU memory, latency, throughput, and cost. How do you choose the right models, deploy them on Kubernetes, and get the infrastructure you need as your application grows?

Overview

Join experts from Google Cloud and NVIDIA at Kube

Con for a developer community meetup and reception focused on building agentic AI with open models. Bring your questions, compare approaches with fellow developers, and explore how to take open models, including NVIDIA Nemotron, from model weights to running applications on NVIDIA-accelerated infrastructure on Google Cloud.

We’ll explore model selection, serving on Google Kubernetes Engine (GKE), and matching infrastructure to your application’s memory, latency, and throughput requirements. Using Google Cloud’s Des