AI & Machine Learning
ML pipelines for churn prediction, forecasting and customer segmentation, built with neural networks, ensemble models and Azure AutoML on real business data.
AI & Full Stack
Software Engineer.
I design and build AI-powered applications, agentic RAG systems, data platforms and scalable full-stack products with Azure, LLMs and modern web technologies. Based in Italy, I usually work with international teams.
AI Engineering
Agentic RAG
Azure Cloud
Full Stack
Data Science
R&D
LangGraph
01About
Short Bio
I'm an AI and Full Stack Software Engineer. I mostly work on proofs of concept, data integration systems and products that sit somewhere between research and production.
Currently at Sintra Digital Business in Arezzo, Italy, I take AI products from the first idea through prototyping to production on Azure. I also mentor students and lecture on deep learning and AI at Italian universities and Microsoft HQ Rome.
4+
Years in R&D
110
MSc grade / 110
3+
Universities as lecturer
02Expertise
Skills
Six areas I work across, from research to shipped product.
Generative AI & Agents
Private LLMs, agentic RAG and AI agents for real enterprise knowledge workflows. I spend a lot of time on prompt engineering and making LLMs behave reliably.
Full-Stack Engineering
Scalable cloud-native apps on REST APIs, microservices and event-driven architectures. OAuth, CI/CD and Shopify/B2C integrations.
Cloud & DevOps
Cloud automation platforms on Azure Kubernetes with Key Vault, Service Bus, Container Registry and CI/CD pipeline management.
Data & Analytics
BI dashboards that turn ML outputs into something useful: clustering, RFM, AutoML forecasting, all wired up through Power BI and Azure Functions.
Product & Innovation
Fast proofs of concept, enterprise consulting for Prada S.p.A., B2C and Shopify integrations, and plenty of experimentation with AI coding tools like Cursor.
03Experience
AI & ML Solutions
Churn prediction, sales forecasting and customer segmentation pipelines built with neural networks, ensemble models and Azure AutoML, often on messy, imbalanced datasets.
Agentic RAG Systems
Private LLM setups with agentic RAG and AI agents using Azure AI Foundry, Microsoft Copilot Studio and LangGraph, built for real enterprise knowledge workflows.
BI & Analytics Dashboards
Power BI dashboards backed by Azure Functions that surface ML outputs like clustering, RFM and forecasting to support day-to-day business decisions.
Cloud-Native Applications
Scalable web applications in .NET, React, Vue.js and Node.js, using microservices, OAuth, CI/CD pipelines and Shopify/B2C integrations.
Cloud Automation Platform
A deployment and infrastructure management platform on Azure Kubernetes with CI/CD, Key Vault secrets and container registry workflows.
Consulting & Lecturing
IT consulting for Prada S.p.A., plus thesis supervision and lectures on deep learning at the University of Rome Tor Vergata, the University of Pisa and Microsoft HQ Rome.
04Projects
Selected work.
Tap any project to see what I actually did.
01
Agentic RAG Platform
Private LLM architecture with retrieval, orchestration and AI agents for enterprise knowledge management
Problem solved
Enterprise knowledge was scattered and unstructured, so teams couldn't query internal documents reliably.
Technical impact
End-to-end agentic pipeline: ingestion → vector retrieval → multi-step LLM orchestration → auditable, grounded outputs.
02
Churn Prediction Pipeline
ML pipeline for highly imbalanced datasets with neural networks, ensemble models and feature selection
Problem solved
Standard classifiers defaulted to majority class on imbalanced data, missing actual churners with high business cost.
Technical impact
SMOTE resampling, covariance-driven feature selection and a stacked ensemble, with churn scores feeding straight into live BI dashboards.
03
Forecasting & Segmentation Analytics
Azure AutoML forecasting, RFM analysis and clustering surfaced in interactive Power BI dashboards
Problem solved
Fragmented customer data meant business teams couldn't see customer value distribution or future demand clearly.
Technical impact
RFM matrix, k-means clustering and AutoML forecasts updated automatically via Azure Functions and surfaced in real-time BI.
04
Cloud Automation Platform
Deployment and infrastructure management on Azure Kubernetes with CI/CD and Key Vault
Problem solved
Manual, error-prone Azure deployments across multiple environments with no centralized secret management.
Technical impact
GitOps workflows, Key Vault-integrated secret rotation, zero-downtime deployments and full observability.
05
Full-Stack Enterprise Applications
Cloud-native apps built on microservices, OAuth and event-driven pipelines, including consulting work for Prada S.p.A.
Problem solved
Legacy monoliths with fragile deployments, disconnected frontend/backend and no cloud scalability path.
Technical impact
Modular architectures with Shopify/B2C integrations, OAuth-based auth and automated CI/CD. Consulting for Prada S.p.A. (.NET, Angular).
05Stack
Technical stack.
Programming
PythonTypeScript
JavaScriptC#
PHPSQL
BashHTML
AI / ML
TensorFlowPyTorch
Scikit-LearnLangChain
TransformersKeras
PandasNumPyOpenCV
Backend & Frontend
.NETNode.js
React.jsVue.js
NestJSExpressFastAPI
Cloud & DevOps
AzureKubernetes
DockerAzure Functions
Service BusAI Foundry
Key VaultCI/CD
Databases
PostgreSQLMySQL
MongoDBAzure Storage
Auth0Pimcore
BI & Analytics
Power BIGoogle Data Studio
Azure AutoMLRFM Analysis
06Education
Academic background.
Master's Degree
Computer Engineering
Università degli Studi di Firenze
SAR image despeckling by means of deep learning methods.
Executive Master
Data Science, Big Data & AI for Finance
24ORE Business School
Stock price prediction using ARIMA models and dense neural networks on financial historical series.
ARIMACRISP-DMFinance ML
Course
Business Plan: Translating an Idea into an Entrepreneurial Project
SDA Bocconi — School of Management
07Knowledge Sharing
From engineering
to the classroom.
I turn hands-on R&D experience into academic content: supervising theses and giving lectures on deep learning, LLMs and agentic solutions.
University of Rome Tor Vergata
Guest lecturer — Deep Learning & AI Applications
University of Pisa
Guest lecturer — Deep Learning & AI Applications
Microsoft Headquarters — Rome
Speaker — AI & Large Language Models
Thesis Supervision
Topics: Forecasting · RPA · LLM · Agentic Solutions
08Beyond Code
Creativity, the most important skill.
To me, engineering and creativity run on the same instincts: structure, iteration and knowing what to leave out. Music and photography shape how I think about system design more than you'd expect.
Blame Rec · Techno / Ambient
Electronic Music
I produce techno and ambient. I'm co-owner of Blame Rec, a record label promoting creative expression for new artists.
Landscape Photography
I photograph natural environments, which keeps me tuned into composition, patience and small details.
CrossFit & Physical Training
I've been training for 8+ years. Regular CrossFit practitioner. PADI Advanced Open Water Diver since 2009.
Open to opportunities
Need an AI engineer who can turn an idea into a working system?
I help teams go from a rough concept to a production-ready architecture, building software that is smart, scalable and actually reliable.
09Contact
Let's build
something great.
New project, consulting gig, or just want to talk? I'm always open.