About Me

I am a data scientist currently working at the Department of Computer Science, Universitat Politècnica Catalunya-Barcelona Tech (UPC). I am a member of the PADS-UPC group.
My interests include any subject relating to data science, with particular emphasis on the implementation of intelligent methods for real industrial issues to automate decision-making processes, and making them interpretable.
I am currently a PhD candidate in Computing for the UPC.

Contact Details

Bernat Coma-Puig
Barcelona
34 93 4137862
bcoma@cs.upc.edu

Work

UPC - Naturgy

Data Scientist 2013 - Present

My research is part of the collaboration project between the UPC and the utility company Naturgy with the aim of detecting fraud customers and other Non-Technical Loss in both electricity and gas using machine learning techniques.
My work focuses on the development of a framework that generates supervised models that detects fraud automatically based on historical cases, delving into the solution of typical problems such as the dataset shift, the lack of quality information, the exploration vs. exploitation trade-off or the explainability of very complex models, without forgetting all the problems that arises from the development of a software that mines teras of data such as the efficient use of the available resources or the multiprocessing.

Education

Universitat Politècnica de Catalunya

Ph.D Candidate 2016-2019

Ph.D in Computing. "A Flexible and Automatic Framework for Fraud Detection in Energy Consumption".

Leibniz Universität Hannover

Computer Science Thesis 2013

Computer Science Thesis (Master Thesis) in L3S Research Center, Leibniz Universität Hannover: "Collaborative Filtering Ensemble for Personalized Name Recommendation"

Universitat Politècnica de Catalunya

Computer Science 2013

Computer Science (5 years Engineer's Degree) at the Universitat Politècnica de Catalunya