I’m an Applied AI Engineer at CSI Piemonte, one of Italy’s largest regional in-house IT providers. I build LLM systems deployed at national scale for Italian Public Administration, a setting where safety and governance matter as much as the machine learning itself.
My current work centers on a tender-assistance chatbot called Camilla. I designed and maintain its semantic search engine (100K+ queries in six months), its behavioral evaluation framework (200+ adversarial test scenarios, LLM-as-a-judge), and the internal MCP server that composes its agent tools. Beyond Camilla, I design agentic assistants for other clients, both inside and outside the company, owning the full stack: specs, knowledge-base definition, and testing of the finished product. I try to build software that stays cost-aware and scalable in production, and together with my team we work to keep it compliant with the EU AI Act and GDPR. I’m drawn to hybrid systems (classical + AI) for the way they combine the performance and explainability of deterministic frameworks with the flexibility, expressiveness, and autonomy of frontier language models.
Alongside this I’m completing an MSc in Computer Science (ML specialization) at Georgia Tech, holding a 4.0/4.0 GPA across courses spanning machine learning, classical AI, NLP, and deep learning. It’s my second master’s: the first was in Mathematics, in Turin, finished with top marks and honors and a specialization in stochastic processes. Math remains my first love, and it’s anything but distant from the job: probability and stochastic processes are the language generative models are written in, and coming at AI from both sides, the mathematical and the computational, gives me the rigor I try to bring to every system I build and evaluate.
Before AI, I spent several years as a business analyst on IT systems handling European funds, from EU agricultural funds (EAFRD) to NextGenerationEU: advanced PL/SQL and SQL, reports and dashboards for officials and EU regulators, where a wrong number had real consequences. It wasn’t AI, but it’s the root of the deterministic half of what I do today, and the reason I believe in hybrid systems: I know what deterministic logic does well before I hand the rest to an LLM. Working alongside European institutions, with their annual audits, taught me how much explainability matters in any system whose decisions carry real weight.
On a personal note: I hold a diploma in classical guitar from the Conservatorio G. Verdi in Turin, where I studied for eight years. I’m passionate about music, mathematics, and artificial intelligence.