Serving LLMs Isn't Serving Web Apps: A Production Kubernetes Playbook
How AI infrastructure evolved from raw GPUs to vLLM, LLM-D, and AI gateways — and how a self-hosted model actually wires into K Gateway on production Kubernetes.
Blog
Thoughts on DevOps, cloud architecture, security, AI, and automation.
What AWS Doesn't Tell You About Building AI Agents on Connect
8 undocumented gotchas from building a production agentic helpdesk on Amazon Connect — circular references, silent failures, missing APIs, and the infrastructure plumbing that ate 60% of the build time.
Beyond 'Smart': A New Mental Model for the AI Workforce
Moving past IQ benchmarks to understand the three axes that actually matter when deploying AI agents: Capability, Autonomy, and Authority.
Terraform Modules The Right Way
Building battle tested terraform modules
Running AI Agents Safely: Building Isolated Sandboxes on Kubernetes
How we solved the multi-tenant code execution problem for autonomous AI agents using Kubernetes, gVisor, and dynamic volume isolation.