Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore the critical infrastructure decisions for deploying AI agents in production through this 41-minute podcast episode featuring Google Cloud experts discussing the technical architecture required for secure, scalable agent deployment. Learn to navigate the serverless versus Kubernetes debate for AI agent runtimes, with specific focus on Google Kubernetes Engine (GKE) capabilities. Discover how to safely execute LLM-generated code without compromising infrastructure security through agent sandboxing techniques using gVisor isolation. Master the implementation of Pod snapshotting technology that dramatically reduces startup latency from seconds to milliseconds, enabling more efficient agent scaling. Follow along with practical demonstrations of deploying Agent Development Kit (ADK) agents on GKE, configuring Kubernetes YAML files, and implementing security measures for code-executing agents. Understand the technical trade-offs between Cloud Run's simplicity and GKE's control for different agent use cases, while exploring advanced features like Pod snapshots that extend beyond sandbox applications to improve overall Kubernetes workload efficiency.
