Articles
Insights from the field
Hands-on writing on platform engineering, AI infrastructure and security, from systems we actually built.
-
Giving AI agents a seatbelt via OpenShell
Confining coding agents with NVIDIA OpenShell: one YAML policy for files, hosts, HTTP paths, MCP tools and where each credential may go, enforced at the kernel and network layers. The short version of a 19-lab course.
-
Agentic SecOps on one workstation: rebuilding the SUSE AI Factory + NVIDIA stack with open source, then testing what it promises
Rebuilding the SUSE AI Factory and NVIDIA stack with open source, then testing seven of its security claims, from deny-by-default egress to audit logging.
-
Built a 5-Node AI Agent in a Weekend — Here’s What LangGraph Actually Teaches You
What LangGraph actually teaches you about agent design, with code and container images.
-
Wired Langfuse observability into a RAG pipeline
Tracing and measuring a retrieval-augmented generation pipeline in production with Langfuse.
-
MLOps vs LLMOps: Why Large Language Models Force Us to Rethink Production AI
Why large language models force us to rethink production AI: tokens, prompts, evaluation and safety after six years of classical MLOps.
-
K3D in A Docker Sample App and Helm Chart
A sample app and Helm chart for Kubernetes-in-Docker test environments.
-
If RSA is not enough for Rapid7 InsightVM/Nexpose Console
Switching the InsightVM/Nexpose console to an elliptic-curve TLS certificate, step by step.