Intelligent Model Routing, Governed: NVIDIA NeMo Switchyard + Kong AI Gateway

About this event

Not every step in an AI agent workflow needs the most powerful and expensive model. So can you ensure every request reaches the right model without sacrificing performance, security, or control? As enterprises move toward multi-model agentic AI, intelligent model routing is becoming critical infrastructure. Choosing the right model is only half the challenge. Organizations also need to secure credentials, enforce governance policies, protect sensitive data, and maintain visibility across every AI interaction. Join experts from Kong and NVIDIA to explore how Kong AI Gateway and NVIDIA NeMo Switchyard work together to make intelligent model routing secure, cost-efficient, and production-ready. Discover how Kong AI Gateway provides centralized security, governance, and observability across AI traffic while NeMo Switchyard dynamically selects models based on quality, cost, latency, and confidence. See the integration in action through a live demonstration and explore benchmark results showing how optimized routing reduced cost per completed task while maintaining task completion rates. You'll leave with: • A framework for secure AI governance: Discover how Kong AI Gateway centralizes credentials, enforces guardrails and PII masking, applies rate limits, and provides auditability across model interactions. • A smarter approach to model selection: Learn how NVIDIA NeMo Switchyard dynamically routes agent requests across open and proprietary models, including NVIDIA Nemotron, to optimize cost and performance. • Real-world performance insights: Explore benchmark results demonstrating how intelligent routing optimization reduced cost per completed task by nearly half, without compromising task completion. • A practical path to production-ready agentic AI: Understand how open model routing and centralized governance support secure, scalable AI infrastructure aligned with the Open Secure AI Alliance's principles.