I'm an AI Solutions Engineer who builds AI agents that do real engineering work — not chatbots that answer questions. I work end-to-end: from understanding a messy problem to shipping a system in production that people actually rely on. Most recently I designed and built an autonomous multi-agent engineering system — an orchestrator that triages tickets and plans work, coding agents that implement in isolated branches, senior-review loops, and human approval kept at the decisions that matter. It's wired across GitHub, ticketing, CI, and team chat: it opens PRs, generates tests, enforces review gates, and asks for approval before touching anything sensitive. Before that, I shipped production AI in regulated settings: a medical computer-vision pipeline serving 45+ clinics, a multimodal conversational AI for Alzheimer's care, and a deterministic RAG assistant for a 210-employee fintech (99% accuracy, zero hallucinations). I work in Python and TypeScript, with FastAPI, LLM APIs, OpenCV, and Azure — and I build *with* coding agents (Claude Code, Cursor, Codex) as primitives I orchestrate, not just tools I use. What makes me stand out: I turn unclear requirements into working systems fast, I've shipped AI where reliability is non-negotiable (healthcare, fintech), and I act as the validation layer that keeps AI output trustworthy — because knowing when the AI is wrong is the real engineering. I also work natively across four languages (Italian, English, Spanish, French), which makes me easy to drop into European and cross-market projects. If you need agentic workflows, AI automation, RAG, or production AI shipped end-to-end, that's exactly my zone.…I'm an AI Solutions Engineer who builds AI agents that do real engineering work — not chatbots that answer questions. I work end-to-end: from understanding a messy problem to shipping a system in production that people actually rely on. Most recently I designed and built an autonomous multi-agent engineering system — an orchestrator that triages tickets and plans work, coding agents that implement in isolated branches, senior-review loops, and human approval kept at the decisions that matter. It's wired across GitHub, ticketing, CI, and team chat: it opens PRs, generates tests, enforces review gates, and asks for approval before touching anything sensitive. Before that, I shipped production AI in regulated settings: a medical computer-vision pipeline serving 45+ clinics, a multimodal conversational AI for Alzheimer's care, and a deterministic RAG assistant for a 210-employee fintech (99% accuracy, zero hallucinations). I work in Python and TypeScript, with FastAPI, LLM APIs, OpenCV, and Azure — and I build *with* coding agents (Claude Code, Cursor, Codex) as primitives I orchestrate, not just tools I use. What makes me stand out: I turn unclear requirements into working systems fast, I've shipped AI where reliability is non-negotiable (healthcare, fintech), and I act as the validation layer that keeps AI output trustworthy — because knowing when the AI is wrong is the real engineering. I also work natively across four languages (Italian, English, Spanish, French), which makes me easy to drop into European and cross-market projects. If you need agentic workflows, AI automation, RAG, or production AI shipped end-to-end, that's exactly my zone.WWWW…