Lightning Talks #4 - AI Needs Infrastructure: What CNCF Is Building

Everyone is talking about AI models. But who runs them in production? Where do they live? How do they scale?

In this session, we explore how the cloud native ecosystem is quietly becoming the foundation for AI infrastructure, and why the shift from model training to model inference is the next big infrastructure challenge.

We will walk through the CNCF projects that are stepping up to solve this: from Kubernetes extensions for AI workloads, to llm-d (a brand new CNCF sandbox project built for distributed LLM inference), to observability tools tracking AI-specific metrics in production.

No AI background needed. If you are curious about cloud native or Kubernetes, this talk is for you.

What we will cover:

• Why inference is replacing training as the core infrastructure problem

• Key CNCF tools running AI workloads today

• What this means for engineers who want to work in this space

Join the CNCF Jordan community for another evening of learning, discussion, and connection.