External Technical Director · Software & AI Systems Architect · Electronic Engineer
I help organizations modernize production software and introduce governed AI workflows without losing control of architecture, data, security or accountability.
- Set technical direction, architecture, constraints and priorities for custom software projects
- Assess, integrate and govern AI-assisted workflows, knowledge bases, RAG systems and agents
- Modernize existing software while protecting security and operational continuity
- Coordinate internal teams and suppliers around clear technical decisions and responsibilities
- Start from operational constraints, ownership and failure modes—not technology fashion
- Keep architecture decisions explicit, testable and reversible
- Introduce AI behind policy gates, observable boundaries and human approval where consequences matter
- Prefer incremental modernization over high-risk rewrites
My hands-on background includes PHP/MySQL applications, server-rendered platforms, backend tooling, Linux production environments, hardening and performance, and mixed software/electronics systems.
Pair demonstrates my approach to maintainable software: explicit architecture, predictable component boundaries, progressive enhancement, conservative migrations and changes that remain reviewable by people and AI-assisted workflows.
ComposerGlass Engine is an independent Composer-compatible dependency engine written in Swift for native developer tools, designed around deterministic behavior, security boundaries, transactional operations and rollback.
An AI-assisted public reference implementation for grounded agents with record authorization, policy gates, persistent approval, audit events and measurable evaluations. It is an inspectable blueprint, not a deployed client system.
- Pair has been developed in public since 2017 and is maintained as a versioned production framework
- Pair 4 uses CI, explicit response contracts, conservative migration tools and private security reporting
- ComposerGlass Engine isolates downloaded code, validates archives and activates dependency trees transactionally
- The Governed AI Agent Blueprint makes AI control boundaries executable and testable rather than leaving them as policy prose
Full professional profile, services and contact form: vmingegneria.it
Based in Crotone, Italy · VM Ingegneria