CTO · AI Innovator · Fintech SaaS Architect
Building bridges between technology vision and business impact. Nearly 20 years at the intersection of architecture, innovation, and delivery.
I design resilient architectures and lead high-performing teams to turn complex challenges into scalable, future-proof solutions.
With nearly 20 years in the Microsoft ecosystem, my journey evolved from deep-dive technical development to leading cross-functional innovation areas spanning Power Platform, Azure, and custom development.
Today I drive digital transformation initiatives for enterprise clients, overseeing AI-powered products including AI4Teams, AI4Align, and AI4Orchestration. My work connects architecture, strategy, and delivery.
Enterprise clients I've worked with include PwC, Medtronic, MPS, and Nexive — across banking, healthcare, logistics, and consulting.
Currently building ToroPulse — smart alerts, deep copier analytics, and social automation for eToro traders.
Lead Architect • 2026 - Present
Your eToro account, supercharged. ToroPulse is an independent research project bringing smart compound alerts (RSI, EMA, Volume), deep copier analytics, and social automation to retail traders and Popular Investors.
Creator • 2026 - Present
AI-generated financial podcast in Italian. A fully autonomous Python pipeline in the cloud: 17 RSS feeds ingested, refined by 3 specialized AI agents for news curation, fact-checking, and editorial tone, then turned into an 8-section script, synthesized via TTS, mixed, and published to Spotify every morning at 07:00.
An AI-powered hub designed to enhance team collaboration, alignment, and productivity within the Microsoft 365 ecosystem.
Intelligent orchestration platform on the Microsoft stack, enabling enterprise-grade workflow automation and AI-driven process management at scale.
Fully local AI pipeline for transcribing meetings, identifying speakers, and generating executive summaries. Works with any platform — Teams, Zoom, Webex. No cloud, no subscriptions, no data leaving your machine.
Five episodes, one full story: a vault where every piece of information has one home, an AI librarian that follows a written contract instead of a prompt, and why context engineering beats prompt engineering. Includes the full technical breakdown.
Headlines say Microsoft lost the exclusive deal with OpenAI. The financial reality tells a different story — Azure stays primary cloud through 2032, $250B in compute, and a stake worth $228B heading into the IPO.
Copilot didn't work for my meetings — not because it's bad, but because my meetings don't follow its rules. So I built something that works regardless of who you invite, what platform, what license.
AI data centers could reach 12% of US electricity by 2028. When institutional investors step back, it doesn't mean AI doesn't work — it means the real bill is arriving. Energy, water, grid — not just GPUs.
That's 7× The Lord of the Rings. In 30 days. Not experiments — real projects, real architectures. A powerful tool used without context produces fast code and slow problems.
Both things are true. Markets are pricing AI as infrastructure. But there's a massive gap between those who build the infrastructure and those who use it to solve real business problems. That gap is my job.
Hands on keyboard, real problem, limited time. We built an AI agent for tracking professional certifications — the kind of problem that runs on Excel and email today, and on an agent tomorrow.
The problem was fully defined. Fully predictable. Fully deterministic. No ambiguity. The right answer wasn't an LLM — it was a switch. Knowing when not to use AI is part of the job.
The second person always wins. Not because they're right — but because the client perceives one thing clearly: this person actually understood what I want. At that point, you're no longer the vendor being evaluated. You're the reference.
Whether you're looking for architectural guidance, exploring AI implementation, or just want to discuss the future of the cloud ecosystem — my inbox is always open.