Platform scope
AWS foundations, Terraform, Kubernetes, CI/CD, runtime governance, observability, cost and security awareness, and the operational contracts that let teams build on the same substrate.
Nick Demming
Notes on platform engineering leadership, AI infrastructure, MLOps, cloud systems, data engineering, and the production boundary where models, services, infrastructure, and teams have to work together.
AWS · AI Platforms · MLOps · Terraform · Kubernetes · Data Engineering · Java · Rust · C++ · Python · Go
This is the revived demming.dev archive and the base for current professional writing. The old Hakyll blog remains preserved, and the new AI maturity package is the first current series: a platform-engineering view of agentic workflows, repository readiness, CI guardrails, and safe delegation.
The through-line is still the same: abstractions should survive contact with production. Whether the subject is Haskell, JVM services, TypeScript serverless systems, Rust components, Kubernetes, Terraform, or model deployment, the interesting work is making the system understandable, operable, and hard to misuse.
AWS foundations, Terraform, Kubernetes, CI/CD, runtime governance, observability, cost and security awareness, and the operational contracts that let teams build on the same substrate.
AI platform engineering, MLOps, evaluation and release gates, model-serving paths, data lineage, and feedback loops that turn useful experiments into maintainable systems.
JVM, TypeScript, Rust, C++, Python, Go, and typed functional ideas applied where interfaces, latency, correctness, and maintenance costs matter.
10 May 2026
A repository is not agent-ready because it has an instruction file. It is agent-ready when context, specs, verification, and boundaries make supervised delegation reliable.
10 May 2026
A practical checklist for making a repository safe for human-supervised AI agents without confusing automation with autonomy.
10 May 2026
AI adoption phases and AI maturity levels are related, but not the same: phases describe transformation sequence, while levels describe proven capability.
10 May 2026
A maturity model for agentic engineering should measure how safely work can be delegated, not how many people have adopted a model or assistant.
10 May 2026
In agentic engineering, CI is the enforcement kernel: agents propose, but tests, policy, review, and humans decide what receives authority.