Writing from the team that also builds the platforms.
One practical write-up per month. Writing about what we learned, what we should have known, and what we'd do differently next time.
Recent pieces.
LLM production readiness for EU enterprise teams.
Most LLM pilots succeed. Most LLM production deployments fail. The difference lives in four layers pilots never test.
READ →Tagging debt compounds. Start with the policy, not the tags.
Every untagged resource today costs three times to fix. Teams that enforce at creation spend three seconds. Teams that tag retroactively spend hours.
READ →Internal developer platform: build, buy, or assemble.
Most teams asking whether to build or buy an IDP are asking too early. There is a prior question that determines the answer.
READ →Kubernetes cost visibility is an engineering problem.
Your cluster runs at 68% CPU. The FinOps review says €180,000 a month. Neither number tells you which team owns which slice.
READ →Commitment discounts: the math most teams get wrong.
The wrong commitment strategy costs an average of 11 percent of the compute bill. The math starts with one different question.
READ →Platform engineering versus DevOps is the wrong question.
Platform engineering is what DevOps looks like when it works at scale. Most teams asking the question are already on the path.
READ →