That's me


About me

I am a machine learning specialist working for Top Network S.p.A. in Rome, Italy, focusing on Deep Learning applied to satellite Multi Spectral Imagery (MSI) as well as other topics in AI such as AI and classical ML.

In the past, I worked for the Hebrew University of Jerusalem on Machine Learning applications to astrophysical data (2020-2021), as a postdoctoral fellow at the Astrophysikalisches Institut Potsdam (AIP) in Germany, between 2018 and 2020 and at the Racah Institute of Physics, Hebrew University of Jerusalem in Israel, between October 2014 and October 2017.

I obtained my Ph.D from the Universidad Autonoma de Madrid in Spain in May 2014, where I also worked as a teaching assistant between 2010 and 2013.

I graduated in Physics (BSc) and Theoretical Physics (MSc) at the Universita' degli Studi Roma 3, Italy. I have been a visiting graduate student at the Heidelberg University, Germany and the Syndey University, Australia. I was a member of the CLUES collaboration, working on the reconstruction of the Initial Conditions of the local Universe using galaxy peculiar velocity data.

For more informations please refer to my CV, the projects page of this webpage, my GitHub repositories, or my Google Scholar Profile.

I am also the founder, manager and bass-player of the parody metal band Nanowar Of Steel, which despite starting as a hobby back in 2003 has now become a pretty serious endeavour, with a solid international fan-base in the tens of thousands-strong and millions of views on YouTube.





A cosmological simulation snapshot


Projects

At the moment I am working on several projects involving the use of multi spectral satellite imagery for the monitoring and prevention of natural disasters, combining meteorological sources with Sentinel / LANDSAT / Modis data. These data are processed using standard statistical techniques as well as ML/AI algorithms to produce risk maps for wildfires and floods. The code for these projects is available on GitHub

In the past years, I have been working on several different topics in cosmology and astrophysics, focusing on the computational and numerical aspects of the problems:

Machine Learning applied to cosmological simulations, developing MLEKO (Machine Learning Environment for KOsmology), a code that applies popular ML libraries (scikit-learn, TensorFlow, PyTorch) to analyze astrophysical data. In particular, I have used supervised learning methods (decision trees, random forest, gradient boosted trees, artificial neural networks...) to infer dynamical properties of the Local Group of galaxies from simulations. I have also applied unsupervised learning algorithms (k-means and Self Organizing Maps) for parameter-free classification in the cosmic web, and convolutional neural networks for pattern detection in the simulations' initial conditions.

Constrained simulations of the local Universe, working on the generation of constrained Initial Conditions with the software tools Ginnungagap and IceCore (for which I have developed an MPI-parallel version). Within the CLUES collaboration I designed the so-called Local Group Factory, a numerical pipeline to simulate Local Group-like objects within a large scale environment that matches current observations. I am developing and mantaining the code PyRCODIO which I am using for simulation analysis and post-processing.

N-body and hydro simulations of non-standard cosmologies such as Quintessence, Coupled and Vector Dark Energy models. This was the bulk of my PhD Thesis, during the course of which I developed a customized version of the popular GADGET-2 code. (PhD Thesis)

Parameter space constraints of Coupled Neutrino-Dark Energy models using the Boltzmann code CMBEASY and MonteCarlo Markov Chain methods. (Master Thesis)





Another cosmological snaphot picture


Resources

Below you can find a selected collection of presentations, articles and videos mainly related to my activity as an astrophysicist, including technical as well as outreach material for the general public. A playlist with most of the interviews I gave for Nanowar Of Steel can be found on YouTube

24/01/2024 From Astronomy to Earth Observations, a lecture on GeoDataScience for the N. Copernicus Observatory in Torun, Poland, entitled Pointing the Telescope in the Wrong Direction (YouTube)

19/07/2023 A primer on the use of satellite data (in Italian) Introduzione pratica all'utilizzo dei Dati Satellitari (YouTube)

03-04/2020 My lecture series for the general public (in Italian) L'Universo Spiegato at Alberto Tomba (YouTube)

06/11/2019 A presentation I gave at the AIP Kafeerunde in Potsdam, Germany on the topic From Astronomy to Music (PDF)

31/10/2019 An article about my research on the Local Group of Galaxies published on Forbes (PDF, Link)

09/07/2018 Nature Astronomy paper I co-authored The Quasi-linear nearby Universe (PDF)

19/05/2016, Seminar at the Beogradska Astronomska Opservatorija in Beograd, Serbia, on the topic Obrnuti inženjering Lokalnog Univerzuma (PDF)

11/05/2016, Presentation for the general public on the subject Galei hakvida ve torat hayakhasut haklalit shel Einstein in Jerusalem, Israel, for the group 20 zot hatkhalah (PDF, Flyer).

22/10/2015, Seminar at Universita' La Sapienza, Roma, Italy, on the topic Reverse engineering the (local) Universe (PDF)

22/08/2014, Invited talk at the CLUES Meeting, Potsdam, Germany, on the topic The MergerTree MPI Code (PDF)

05/09/2013, Contributed talk at the PACIFIC 2013 Symposium, Moorea, French Polinesia, on the topic The imprints of quintessence dark energy on the cosmic web and galaxy clusters (PDF)

22/12/2011, Interview on The New Scientist about the influence of dark energy on old massive galaxy clusters (Link, PDF)

02/09/2011, Contributed talk at the Azores Summer School on Observational Cosmology, Angra do Heroismo, Portugal, on the topic Massive high-z clusters and alternative cosmological models (PDF)





L'Universo Spiegato ad Alberto Tomba


Podcast

I am running a weekly show called the GattoCast, where I mainly discuss music, languages and science, mostly in English but often also in Italian and other languages. You can find it on all major podcasting platforms such as:





A Knight at the Opera by Nanowar Of Steel


Extra

My main activity outside of science and programming is related to Nanowar Of Steel, which is started in 2003 and counts hundres of shows in Italy and Europe, five albums, a live DVD, thousands of copies sold, 15+ million of streams on Spotify and 30+ millions of views on YouTube; we are currently signed to Napalm Records. You can check us out on YouTube, on Facebook, on Instagram and on Spotify.

I also like foreign languages, besides Italian and English I am fluent in Spanish and Serbo-Croatian; I have a good command of German, French, Russian, modern Hebrew and Arabic (mainly Levantine dialect but also MSA). I can have meaningful conversations in Portuguese and at the moment I am learning Mandarin Chinese.

In what's left of my spare time, I enjoy reading as much as possible about Machine Learning and Artificial Intelligence, Economics and Finance, History (breakup of Jugoslavija, former Soviet Union and modern Middle East in particular) and Russian literature (Dostoevskij, Gogol', Tolstoj ...).

My philosophical and religious views are summarized in this paper: Foundations of Gauge Theology.





Get in touch!


Contacts

If for any reason whatsoever you feel compelled to get in touch with me please do so using any of the links you can find on the navigation bar of this page.

Good old fashioned email is still the best option - you can drop me a line at:

gatto @ nanowar . it

Please specify whether you're getting in touch with me through this site, just so I know someone is having a look at it at all.

By the way - you can probably tell from the low quality of these pages that I wrote the HTML/CSS code myself. I did it on a Linux (Kubuntu) terminal using what is by far the best IDE/Text Editor on Earth : Vim
(I know it's technically not an IDE, but still)

This website is proudly hosted (as you might have guessed) on GitHub, which is most likely the only Microsoft-related tool I've used in the last decade.