My name is Marcelo Prates. I hold a PhD in Computer Science with Major in Machine Learning from UFRGS. During my doctorate (Aug 2015 - Aug 2019) I’ve studied Neural-Symbolic Reasoning / Graph Neural Networks (i.e. training deep learning models to solve symbolic problems) and the Ethics of Artificial Intelligence, in particular Machine Bias.
I currently (March 2020 - Today) work as an AI Researcher at Samsung R&D Institute Brazil. I was previously (Aug 2019 - March 2020) a Data Scientist at Poatek, where I worked on a wide range of interesting projects, including risk analysis , relationship extraction and visualization applied to geospatial data. Many of these projects included machine learning or deep learning to some extent (convolutional neural networks, gradient boosting / ensemble methods and pre-trained language models).
I also work as a Volunteer Mathematics Teacher at ONGEP (April 2018 - Today). ONGEP (Non-Governmental Organization for the Popular Education) is a non-profit preparatory course for University admission exams in Brazil. Its goal is to provide high quality education at an expendable price for low-resource students. All teachers and collaborators are volunteers, which enables ONGEP to charge students only about R$700 a year. It has been a rewarding experience both from the personal and professional perspectives. It is a challenge to cover so many topics in so little time and to present them in a way that’s accessible to students with gaps in their formal STEM education.
Finally, I’ve been playing around with Generative Art for some time now (2015 - Today). My main interests include abstract artworks and the interplay between visual arts, mathematics and the natural sciences – in particular with respect to creating art through computer simulations of natural phenomena. You can see a portfolio of my projects (such as the one featured below) in the Portfolio Tab.