I am a third-year PhD Student in Statistics & Machine Learning at the Department of Statistical Science, University College London and The Alan Turing Institute, under the supervision of Ioanna Manolopoulou and Gianluca Baio (intake: September 2019). Previously, I worked as researcher and consultant at the The Bartlett School of Environment, Energy and Resources, UCL. I graduated with a Master of Science in Econometrics in 2017 and with a Bachelor of Science in Economics in 2015, both from Bocconi University (Milan, Italy).
My research mainly involves combining the two fields of Machine Learning and Causality, with particular focus on Probabilistic Machine Learning and Bayesian Nonparametric methods.
I am currently working on methodology and extensions of Causal Machine Learning algorithms for Off-Policy Learning and for predicting Individualized Treatment Effects (ITE). Estimation of the effects of an intervention at an individual level is of interest in many disciplines (precision medicine, socio-economic sciences, personalized marketing, etc.) where learning heterogeneous response, in order to design highly-personalized optimal policies, is crucial.
I worked on a few research/consultancy projects commissioned by the UK governmental Department for Business, Energy & Industrial Strategy, performing statistical and econometric evaluation of environmental/energy policies. I have worked (or I am currently working) on a number of consultancy projects involving statistical analysis and App development for customized data analytics (mainly through R Shiny).
I am (or I have been) serving as a Teaching Assistant for the following courses at UCL:
PhD in Statistics & Machine Learning, 2019 - present
University College London
MSc in Statistics & Econometrics, 2015 - 2017
BSc in Economics, 2012 - 2015
This project’s focus is on the use of Bayesian Nonparametric regression models for individual/heterogeneous treatment effects estimation and policy evaluation. In particular, the interest is in Bayesian Regression Trees methods (Bayesian Additive Regression Trees - BART), and their causal version, Bayesian Causal Forests.
In the recent years, there has been a surging interest in the use of Statistical/Machine Learning (ML) tools for causal inference. These tools can leverage large datasets and usually deliver excellent predictive performance.