Hi! I am Alessandro

I research and write about things at the intersection of Computer Science, AI and Society.

Born and raised in Northern Italy, with some more southern roots, I moved to the Netherlands to study and work. In 2026 I saw my first academic paper presented in San Diego. The same week I received an award for a journalism project.

For the remainder of the day, I am a passionate musician and a hiker (the latter not being very easy to practise in the very flat Netherlands).

ACL 2026

GENOME: A new Geopolitical Event Methodology and Dataset using Large Language Models is a study on the use of LLMs to extract cooperative and conflictual events from newswire sources and part of my internship project at The Hague Centre for Strategic Studies (HCSS). Presented at the Association for Computational Linguistics (ACL) annual conference, it received an Honourable Mention at its workshop.

MyLLENNIUM Award

Together with the journalist Gloria Pessina, I wrote an inquiry on the use of AI in the Italian legal sector. We received a prize for the best reportage. The article will be published on Linkiesta in October.

Developer

On top of the research paper, I curated the API, the automations and the documentation related to the GENOME dataset, which HCSS aims to distribute to the public soon.

During my internship, I also curated the refactoring of around 10 different dashboards: standardising and securing the backend, automating data updates from external APIs, and redesigning the UI. Most of this took a lot of time and sweat, and ironically happened a couple of months before the moment when a 20-minute session with Claude Opus could do the same job.

I took it philosophically, as it makes me appreciate the transition from software developer to AI engineer even more.

Academic Background

Chronically indecisive between STEM and Public Policy, I have a background that spans from microelectronics, machine learning and Bayesian statistics to political decision-making and behavioural policy models.

I hold a BSc in Electronics Engineering from Politecnico di Milano, which gave me a highly theoretical foundation. I am not a huge fan of circuits, however.

I am now completing my MSc in Engineering and Policy Analysis at Delft University of Technology. There is still a lot of computer and data science, but it is applied to modelling socio-technical systems, together with some tracks on ethics, behavioural theories, economics and strategy.

Some projects from my uni times

Bangladesh road network model

Agent-Based Modelling for Bangladesh Road Network

In this group project, I replicated a TU Delft research study for the World Bank, aimed at optimizing Bangladesh's road and bridge network. Starting with raw, unstructured data from the Ministry of Transport, I performed extensive data cleaning and exploration to construct a functional network model of the country's major highways.

IJssel River flood risk model

IJssel River Flood Risk Management

In this advanced simulation project at TU Delft, our team acted as policy analysts for the dike rings of Deventer and Gorssel to develop a robust flood risk management strategy for the IJssel River Basin. Operating under conditions of "deep uncertainty," we moved beyond traditional predictive modeling to explore thousands of plausible future scenarios involving climate change and socio-economic volatility.

Mini (Bayesian) GPT

This is a comparative study of tokenization strategies and language model architectures for text generation, built on Andrej Karpathy's miniGPT. We implement three models of increasing complexity (a Neural Bigram Model, a GPT Language Model with multi-head self-attention, and a Monte Carlo Dropout GPT for uncertainty quantification) and evaluate them across three tokenization methods and two datasets.

Spatial data analysis maps and charts

Spatial Data Analysis on India and The Hague

This project applies data cleaning, Global and Local Spatial Autocorrelation (Moran's I and LISA), Principal Component Analysis, and K-Means clustering to different datasets regarding socio-economic patterns in different geographical scopes.

Open Source

I recently built an MCP server that lets LLMs explore Italian legislation natively through Normattiva, the official repository of Italian law. It wraps the IPZS OpenData API as MCP tools, resources, and prompts. In the pictures, a small Sonnet prompt returns a verified overview of all the Italian law acts published in the last month.

The MCP has recently been added to the Italia MCP Servers list.

CV and Contacts

If you are interested in any of my works, want to collaborate or simply have a coffee chat, please contact me through any of these channels: email, LinkedIn or GitHub.

Here you can also find my CV.