on Kubernetes, Serverless, Cloud, Devops, Karpenter, Keda
The promise of Kubernetes was always “deploy anywhere, scale automatically.” The reality was years of kubectl cordon, manual node group sizing, and 3am PagerDuty alerts about cluster autoscaler lag. That gap is closing fast. In 2026, the combination of Karpenter, KEDA, and managed control planes is delivering on the original promise.
on Rust, Backend, Performance, Systems programming, Axum
Rust has crossed the chasm. It’s no longer “that language with a steep learning curve that rewrites Firefox components.” In 2026, Rust is powering production backend services at companies ranging from startups to hyperscalers, and the ecosystem — especially web frameworks — has matured to the point where choosing Rust for a new backend service is a reasonable, defensible decision.
on Platform engineering, Devops, Developer experience, Idp, Backstage
Platform engineering has moved from hype to discipline. In 2026, most companies with more than 50 engineers have a dedicated platform team — or wish they did. But there’s a graveyard of internal developer platforms that nobody uses. Here’s what separates the successful ones from the forgotten portals collecting dust.
on Observability, Opentelemetry, Ebpf, Monitoring, Devops, Profiling
The three pillars of observability — logs, metrics, traces — were the consensus model for the better part of a decade. In 2026, a fourth pillar has emerged: continuous profiling. Combined with eBPF-powered automatic instrumentation and the now-ubiquitous OpenTelemetry standard, modern observability looks quite different from what we built just three years ago.
on Ai, Llm, Fine-tuning, Machine learning, Production
Fine-tuning large language models has matured rapidly. What once required a warehouse of GPUs can now be done on a single A100 with the right techniques. This guide covers what’s actually working in production environments in 2026.
메모는 기억보다 확실하다