I don't rent my AI models
Ollama on-prem on the Hetzner VPS, zero US cloud: I own the infrastructure I use every day, I do not rent it from a third-party provider.
What I believe, and what I will not compromise on.
The position statement: the few things on which I will not compromise.
What I believe
I believe technical depth is proven by a live production system, not with tutorials. Backend, frontend and professional paths have equal dignity: no one front prevails over the others. AI accelerates mechanical work, but I remain the engineer who decides on architecture, security, and trade-offs.
A separate category
Most Italian technical paths stop at one front: backend, frontend, or career. I keep them all on the same proof-of-work: a multi-microservice monorepo that's really in production. The frontend is not a separate course: it's a complementary pillar, tested by the same system and to the same standards as a production-grade backend and solid IT professional paths.
I don't agree with this translation. A more accurate translation would be: What am I not carrying?
Every position below is anchored to a verifiable fact, starting from the lighthouse: owning your own AI models via Ollama on-prem, not renting them from a third-party cloud.
Production-first, never demo-ware
Every technical claim is verifiable on a 3-microservice monorepo actually in production, with architecture defended by automated tests in CI, not on a slide.
AI accelerates, it does not decide
AI writes the mechanical scaffolding. Architecture, security and trade-offs stay my decisions: I remain the accountable engineer.