IASF Milano 
Websites: http://www.mariopasquato.com; http://www.risingframework.eu; https://www.10nebulae.art 
Code: https://gitlab.com/mariomario Kaggle: http://kaggle.com/mariopasquato

Research Experience 
Researcher, IASF Milano 01/03/2024 – permanent 
I use machine learning tools to make great discoveries in the X-ray sky with the EXTraS team, while also pursuing a long term research program in machine learning causality, interpretability and fairness. 
Marie S. Curie Individual Fellow, U. de Montréal & Padua University 01/09/2021 – 30/02/2024 
Project RISING (Realistic and Informative Simulations with machine learnING) received the seal of excellence by the European Commission and was funded with ~250 k$ from Sept. 2021. As a global MSCA fellow I spent two years at Montreal University –one of the leading centers for deep learning development in the world- working on generative deep learning models for astrophysical simulations and on causal discovery applications to cosmology in collaboration with Prof. Yashar Hezaveh and Prof. Yoshua Bengio. 
CAP3 Fellow at New York University Abu Dhabi, United Arab Emirates, 01/09/2020 – 30/08/2021 
Interdisciplinary, independent fellowship aimed at applying machine learning techniques to cosmology (especially zoom simulations), planetary and particle astrophysics
Marie S. Curie COFUND Fellow, Padua Observatory, Italy, 01/09/2017- 30/08/2020 My project was named ARTificial Intelligence Search for Intermediate-mass black holes in star Clusters - ARTIStIC. I worked on a variety of astrophysical problems using deep learning (e.g. on mock images from star cluster simulations) and more conventional machine learning and clustering techniques. Co-supervised six students. 
Senior Data Analyst, Kvadro-M, Ukraine, 01/06/2017 – 31/08/2017 Kvadro-M is a leading wholesaler of dental supplies in Ukraine. The company recently moved into the retail space, opening an e-commerce website (http://kvadro-m.com). I analyzed web marketing performance indicators in R and Tableau, and worked with Google tools, e.g. Search Console, Analytics, Adwords, and 3rd party SEO tools (Semrush). 
Postdoctoral fellow, Yonsei University, Korea, 01/10/2012 – 31/05/2017 I developed my own computational astronomy research program with a focus on star-cluster dynamics and intermediate-mass black holes. Run and analyzed large sets of direct N-body simulations (with the NBODY6 code on GPUs).                                                  
Between 30/06/2012 and 01/10/2012 I took some time off to prepare for moving to Korea, including a period to familiarize myself with the Korean language and culture.        
Postdoctoral researcher, Bologna University, Italy, 01/07/2011 – 30/06/2012 Performed the simulation groundwork for Ferraro et al. 2012, a Nature paper on the formation mechanism of blue straggler stars in GCs. We reproduced their bimodal distribution with mass segregation of a pure binary population. Contract Consultant, Pisa, Italy 27/01/2011 – 30/04/2012 I was registered as a professional in the field of education support activities for the period indicated (overlapping my first postdoc), with VAT number 01994790507. I provided training in statistics and computer skills, either on an individual or company level.  

Education 
Ph.D., Astrophysics, Pisa University, Italy, 01/01/2007 – 07/12/2010 Full, merit-based scholarship and 
tuition waiver. Thesis: Globular clusters and intermediate-mass black holes: a model-free, non-parametric approach, between simulations and observations -  Advisor: Prof. G. Bertin (Milan University). During my Ph.D. course I visited the European Space Agency ESTEC facility (Noordwijk, The Netherlands June 2009; expenses fully paid by ESA) and was a summer intern at Space Telescope Science Institute 
(Baltimore, USA, June-August 2008; expenses fully paid by STScI). 
M.Sc., Physics Scuola Normale Superiore, Italy, 01/10/2001 – 01/10/2006 Full, merit-based scholarship and tuition waiver. Grade 70/70, summa cum laude.  
M.Sc., Astrophysics Pisa University, Italy, 16/10/2004 – 26/09/2006 Grade: 110/110, Thesis: The 
fundamental manifold of globular clusters, Advisor: Prof. G. Bertin (Milan University)
B.Sc., Physics Pisa University, Italy, 12/10/2001 – 15/10/2004 Grade: 110/110, Thesis: Stochastic resonance and its applications to biology, Advisor: Prof. R. Mannella (Pisa University)

Skills and assets  
Programming Languages 
Python – fluent (code sample using Keras: https://gitlab.com/mariomario/jellyfishtng) 
R - fluent (code sample: most repositories in https://gitlab.com/mariomario) 
C – working knowledge (code sample: https://gitlab.com/mariomario/quesodish) 
Machine learning 
Unsupervised learning in R: t-SNE and UMAP for dimensionality reduction (libraries tsne, umap), clustering with DBSCAN (library dbscan) and hierarchical and partitioning methods (library cluster). Supervised learning with traditional (not deep learning) tools in R such as SVM (library e1071), k-nearest neighbor (library FNN), trees and random forests (libraries C5.0, party, randomForest). Neural networks in R using the h2o library. Supervised learning with deep convolutional neural networks on images using Keras on top of Tensorflow in Python. Took part in several competitions on the kaggle.com platform, with results in the top 5% out of thousands of users. Selected competitions: Africa Soil Property Prediction Challenge top 5%; Prudential Life Insurance Assessment top 23%, awarded silver medal for kernels; Driver Telematics Analysis (sponsored by AXA insurance) top 34%; Diabetic Retinopathy Detection top 46%. 
Astronomy specific computing skills 
Codes for star-cluster dynamics: NBODY 6 and MOCCA (I have access to the source code and to the MOCCA Survey Database I library of simulations through my collaboration with Prof. M. Giersz at the Copernicus Center for Astronomy in Warsaw, Poland). Planning and deploying large simulations, including on high-performance computing clusters, using the relevant Unix/GNU- Linux and HPC tools (shell scripting, schedulers, performance monitors...). Writing, debugging and optimizing scripts for data processing, model fitting, plotting, and visualization both in R (using ggplot2) and in Python (using astropy, matplotlib, seaborn). Some experience with SAOImage DS9, topcat, GRAVPOT16. Astronomical database queries using SQL. As part of Prof. Mapelli’s group at Padua University (both former and upcoming member) I have access to her set of thousands direct N-body simulations of star clusters. As part of Prof. Macciò’s group at NYUAD I have access to the NIHAO (Numerical Investigation of a Hundred Astronomical Objects) zoom-in cosmological simulations.