Token Governance, AI Harnesses, and Bare Metal AI Infrastructure at Scale | Rob Hirschfeld, RackN | TFiR
Rob Hirschfeld of RackN explains why AI harnesses must come before model selection, why local GPU workstations ...
Distributed Cloud vs. Centralized Cloud for AI Inference | Ari Weil, Akamai | TFiR
Ari Weil of Akamai breaks down why centralized cloud deployments fail under production AI inference load, how ...
Agentic AI Inference Beyond the Centralized Cloud | Jon Alexander, Akamai | TFiR
Jon Alexander of Akamai explains why centralized cloud compounds latency in agentic AI systems, why up to ...
The Agentic Data Loop: Discover, Read, Write at Scale | Michel Tricot, Airbyte | TFiR
Michel Tricot, Founder and CEO at Airbyte, explains why MCP wrappers fail in production agentic systems, how ...
The RBAC Reality Check for AI in Platform Engineering | Corey McGalliard, Akamai Cloud | TFiR
AI agents in Kubernetes require scoped RBAC and policy as code before write access. Corey McGalliard, Akamai ...
90% of AI Agent Latency Isn’t the GPU | Jon Alexander, Akamai
Agentic AI loops compound latency exponentially in centralized cloud. Jon Alexander of Akamai explains why distributed inference ...
Why Centralized AI Inference Fails at Scale and How Distributed Edge Infrastructure Fixes It | Dr. Robert Blumofe, Akamai | TFiR
Dr. Robert Blumofe, CTO at Akamai, explains why centralized GPU clusters cannot serve ubiquitous AI agent demand, ...
Why Unmanaged AI Agents on Developer Laptops Are an Enterprise Security Crisis | Miska Kaipiainen, Mirantis | TFiR
Miska Kaipiainen of Mirantis Lens explains why AI coding agents running on developer laptops are an immediate ...
How to Run AWS Locally and Cut Cloud Dev Costs | Waldemar Hummer, LocalStack | TFiR
Waldemar Hummer explains how LocalStack runs the cloud locally, why AWS and Docker are partnering with the ...
Why Platform Engineering Teams Over-Abstract and How Modular Design Fixes It | Corey McGalliard, Akamai Cloud | TFiR
Akamai Engineering Manager Corey McGalliard explains why abstracting too much too early breaks internal developer platforms and ...