Build and Run AI Agents at Scale with LangGraph and CubeSandbox on OCI Bare Metal
Workshop Overview
In this hands-on workshop, we'll build a task-oriented AI agent from scratch using the open-source projects LangGraph and CubeSandbox — and deploy it on Oracle Cloud Infrastructure's Bare Metal or your own machines.
Along the way, we'll explore how modern agent harness architectures go beyond LLMs, covering stateful workflows, secure tool execution, sandboxed code runtime, and scalable agent infrastructure.
Whether you're building your first production agent or exploring enterprise agent architectures, you'll leave with a working application and a practical understanding of how to run AI agents reliably at scale.
Speakers
Lillian Shao is Vice President of APAC Business Development at Oracle Cloud Infrastructure (OCI), where she leads strategic partnerships and go-to-market initiatives for the Cloud and AI Infra business.
Lillian has led global expansion and market entry strategies for leading cloud hyperscalers, including OCI and Amazon Web Services (AWS). She works closely with customers, technology partners, and ecosystem leaders to accelerate AI adoption across industries. Lillian holds an MBA from The Wharton School.
Feynman Zhou is the Head of Product for Tencent Cloud Agent Runtime and a CNCF Ambassador. He is an open-source maintainer across several cloud-native projects.
Over the past decade, he has worked at Microsoft Azure, QingCloud, and HPE, building cloud-native platforms and developer infrastructure. Today, he focuses on helping enterprises build and run AI agents without the complexity of managing agent infrastructure themselves.