Lorenzo Liuzzo

Physicist and software engineer specialized in AI. I came to software through physics — derive the model first, then write the code — and I now spend most of my time on how much of that writing can be delegated to a model, and where a person still has to sign off.

Curriculum Vitae (EN) Versione italiana

What I’m building

MyThingsLab
A personal lab for AI-assisted development, started in July 2026: more than 50 small Python tools on one dependency-free SDK, covering research, design, implementation, testing and documentation. Headless workers do the deterministic work themselves and call the model only where judgement is needed, then open a pull request — never a merge, so the last word stays mine. It is an experiment in control, not a product.
Money Balling AI
A team of four building one model that predicts the exact box score of an NBA game — every player and team line — with calibrated uncertainty. A graph encoder over lineup stints, a Monte Carlo possession simulator, and calibration scored against the real box score. I own the architecture.

Interests

AI-assisted software engineering
Where a model earns its place in the loop, and what has to stay deterministic, tested and gated around it.
Scientific machine learning
Applying ML to problems in physics and astronomy, where the model has to respect the physics rather than just fit the data.
Computational physics & simulation
Ray tracing, Monte Carlo methods and optical systems modelling.

Education

  1. 2025 — present

    MSc in Artificial Intelligence for Science and Technology

    University of Milan · Milano-Bicocca · Pavia

  2. 2021 — 2025

    BSc in Physics

    University of Milan 100/110

  3. 2016 — 2021

    Scientific Diploma

    I.I.S. Severi-Correnti, Milan 100/100

Technical Skills

Programming languages

  • Python
  • C++
  • Julia
  • C
  • Nim

AI engineering

  • Claude
  • Claude Code
  • Agentic workflows
  • Prompt & context engineering

Machine learning

  • PyTorch
  • PyTorch Geometric
  • Scikit-learn

Scientific computing

  • NumPy
  • SciPy
  • Pandas
  • Matplotlib
  • Jupyter

Quantum computing

  • PennyLane
  • Qiskit

Databases

  • Neo4j
  • Cypher

Practices

  • Test-driven development
  • Code review
  • CI/CD
  • Static typing

Dev tools

  • Git
  • GitHub
  • GitHub Actions
  • pytest
  • ruff

Typesetting

  • LaTeX
  • Typst
  • Markdown
  • Jekyll

From Friends

Everything above is my own account of myself. This part isn’t: it’s what the people who actually know me have to say, in 140 characters or fewer.

Know me? Leave a line. I read everything before it goes up, so give it a day or two to appear.

Contacts

The best way to reach me is email. I’m also on GitHub, where I host the code behind the coursework and side projects listed under Projects.