Thoughts on platform engineering, AI, automation, and building technology that scales.
The agents started creating scripts to codify their own work before we had named the pattern. Here is what the platform built when it formalised it: certification gates, execution tracking, and the enforcement layer that became the Agent Enforcement Plane.
The agent confirmed the fix. CI passed. The next deploy failed with the same errors. Here is what we learned about determinism, ephemeral clusters, and why prompt templates are not enough.
Policy documents and CI gates are not enough when agents can still act directly. We are building the runtime layer that makes governance operational.
A Gemini agent decided exploratory conversation was a task list and began inserting metadata headers across 700 files. Here's the branch isolation, PR gates, and CODEOWNERS enforcement that stopped it — and the platform architecture that came out of the incident.
Everyone knows Kubernetes runs containers. Almost nobody realises it's actually an extensible API for encoding operational knowledge as code, and that distinction changes how you build platforms.
We didn't choose Kubernetes, ArgoCD, and Backstage because they were free. We chose them because building on open, community-governed tools is the only platform strategy that compounds in your favour.
Why most companies hiring 'platform engineers' are buying the wrong thing, and what the CNCF's hardest exam reveals about the role.
How we're building a governed RAG system with knowledge graphs, hybrid retrieval, and contract-validated outputs, and why we're sharing the journey before reaching production.
How a missed session writer incident led us to build a governance portal with four-layer enforcement, and how the platform started recommending its own improvements.
Three CNCF whitepapers. Three pillars of modern infrastructure. Here's how we unify platform engineering, cloud native AI, and the operator pattern into a single coherent model.
Kubernetes Operators encode operational expertise into automated controllers. The CNCF's Operator whitepaper defines the pattern, capabilities, and best practices for building them.
AI workloads are fundamentally different from traditional cloud native applications. The CNCF's Cloud Native AI whitepaper maps the gaps, and the path forward.
The CNCF TAG App Delivery published a definitive whitepaper on platform engineering. Here's what it says, why it's important, and how it aligns with what we're building at Scaletific.
Internal Developer Platforms aren't just for FAANG. Here's why your growing team needs golden paths, governance, and self-service infrastructure.
