AI agents are everywhere right now, but turning a cool proof-of-concept into something reliable, observable, and scalable in production is still the hard part. In this talk and live demo, we break down what AI agents are, how they work, and why production readiness is the real differentiator.
Seldon’s Paul Bridi and Antanina Vertsinskaya define agentic systems through a practical customer-support example, then unpack the three biggest hurdles teams hit on the road to production: design & build, observability, and deployment at scale. You’ll also see a real-time Kubernetes demo of a marketing AI agent built with Seldon Core and the LLM Module, featuring tool-calling, memory, and an inference pipeline designed to handle real workloads.
If you’re building agents and wondering how to make them production-ready, this session is for you.
