Treat model output as a probabilistic subsystem, not a smarter function call.
AI Inference Architecture: Building Software Around Uncertainty
Razvan Mares··8 min read
Most software systems are built around a comforting assumption: given the same valid input, a component should produce the same correct output.
Advanced AI features break that assumption.
A language...
PostgreSQLci-cdcsharp+2
Turn long-running AI coding sessions into recoverable, reviewable engineering work.
Claude Code Session Design: Checkpoints, Handoffs, and Safe Long Tasks
Razvan Mares··9 min read
Claude Code is most effective when treated less like an autocomplete tool and more like a temporary engineering environment. It can inspect a repository, trace behavior across layers, propose a plan, ...
ci-cdnodejsTypeScript+2
Use Terraform for infrastructure, Ansible for convergence, and k3s for a repeatable platform.
The Control Plane Relay: Provisioning k3s with Terraform and Ansible
Razvan Mares··7 min read
A reliable k3s cluster is not created by one tool doing everything. It is created by a clean handoff between tools with different responsibilities.
Terraform is excellent at describing **what infrast...
linuxproxmoxk3s+1
A practical tour of the small C# 14 changes that make APIs, models, and callbacks easier to write.
Seven C# 14 Features That Remove Friction From Everyday .NET Code
Razvan Mares··8 min read
C# 14 is not a release built around one dramatic new abstraction. It is a collection of language improvements that target familiar sources of friction: extension APIs that feel constrained, properties...
.NETdotnet-10csharp
Where C# 14 removes everyday boilerplate in APIs, validation, spans, and mutation.
A Friction Audit of C# 14: Six Features That Simplify Real Code
Razvan Mares··7 min read
C# 14 is not a release defined by one dramatic new abstraction. Its value is more practical: it trims recurring sources of friction that appear in mature codebases. A conditional update no longer need...
.NETdotnet-10csharp
Treat infrastructure, operating systems, and Kubernetes as distinct stages of one reliable delivery pipeline.
The Cluster Assembly Line: Terraform, Ansible, and k3s Without Drift
Razvan Mares··8 min read
A k3s cluster is deceptively easy to create. One command can turn a Linux machine into a Kubernetes server; another can attach an agent. That convenience is valuable, but it can also hide the engineer...