AI
Building a Two-Node LLM Inference Cluster on NVIDIA GX10 (GB10)
Two GX10 boxes, a 230B MoE coding model, and one flag I should never have set. How 25 tok/s became 40.4, and why it stops there.
• 17 min read
4 posts
Two GX10 boxes, a 230B MoE coding model, and one flag I should never have set. How 25 tok/s became 40.4, and why it stops there.
MCP servers are everywhere, but security is underestimated. Learn how to implement transport encryption, RBAC, audit logging, and credential protection for production AI tooling.
How I built and open-sourced a Model Context Protocol server that enables Claude AI to automate Cisco Nexus Dashboard operations - from architecture to GitHub release.
A network engineer's journey from WordPress frustration to a modern, blazing-fast blog built with Next.js—powered by AI collaboration with Claude Code.