on Llm, Ai engineering, Evaluation, Mlops, Production ai
Shipping LLM-powered features is now table stakes. The hard part isn’t getting a prototype working — it’s knowing whether your model is performing well, catching regressions before users do, and maintaining quality as you iterate on prompts and models.
on Kubernetes, Cloud native, Devops, Platform engineering, Container orchestration
Kubernetes is nearly a decade old now, and the ecosystem around it has accumulated enough churn to confuse even experienced practitioners. Patterns that were best practice in 2022 are anti-patterns in 2026. Tools you invested in have been deprecated, merged, or superseded.
on Gitops, Devops, Argocd, Flux, Kubernetes, Continuous delivery
GitOps has crossed the chasm. What started as a blog post from Weaveworks in 2017 is now the default CD model for Kubernetes-native organizations. The Git repository is the source of truth. The cluster reconciles itself to match. Humans don’t kubectl apply in production.
on Javascript, Node.js, Deno, Bun, Backend development, Runtime
For most of JavaScript’s server-side history, the conversation was simple: you use Node.js. It was the only serious option, and the ecosystem — npm’s millions of packages — made switching cost prohibitive anyway.
on Openclaw, Aws, Bedrock, Claude, Ai
This guide walks through how to connect OpenClaw to AWS Bedrock and run Claude Sonnet 4.6 as your primary model on macOS. Using Bedrock lets you leverage your AWS account’s capacity with stable, reliable inference — and once it’s set up, it just works.
메모는 기억보다 확실하다