Infrastructure, observability, trust / Draft
Systems that can explain themselves
A note on why explainability is not only a model property, but a system property.
Italy / Europe
Alessandro Verrucchi
Systems thinker, AI systems architect, infrastructure engineer, and founder.
I work on resilient software systems, AI-native infrastructure, and the operating principles behind dependable technical organizations.
Signals
A compact view of the work behind the claims.
Practice
Architecture
Designing systems where reliability, cost, latency, and human understanding are treated as linked constraints.
Perspective
Founder
Building from first principles while keeping strategy close to implementation detail.
Focus
AI systems
Exploring how intelligent systems should be deployed, observed, constrained, and maintained.
Recent thinking
Drafts and research directions now shaping the site.
AI systems, economics, architecture / Outline
The cost surface of AI products
How latency, token usage, memory, and quality form a practical design space.
Research
Resource-aware AI systems
How should AI infrastructure reason about cost, latency, memory, energy, and accuracy as first-class constraints?
Working notes and implementation patterns in progress.
Contact
Direct, specific, and low-noise.
Reach out for thoughtful technical conversations, founder-to-founder exchanges, AI systems architecture, infrastructure questions, or speaking invitations.