Diego Spiering Pires

Background

About Diego Spiering Pires

As a physicist: The academic training of a scientist requires a person reach at the frontier of the knowledge in his field and continue to work to push this horizon on unexplored territories. For such achievement, the scientist must remain constantly absorbing and developing new knowledge. So it was my last ten years, dealing with questions which nobody knew the answer, breaking them into a series of particular minor problems, and then spending days/weeks/months applying different techniques to solve every part of the problem. At the end comes the steps of organization, interpretation and divulgation of the final results. The narrative (storytelling) is fundamental because the researcher - author of this new results - needs to convince other scientists about the relevance of his idea so that this idea could be published and then integrated to the standard knowledge of that field. In addition to requiring extensive analytical tooling, this whole process also demands that the scientist see each problem from the most varied angles, always pursuing the most appropriate approach.> As a data scientist: During the last two years, in addition to my academic career in the field of high energy particle physics, I have been dedicating myself more and more to topics related to computer science, probability and statistics. I have been taken some disciplines in these areas at the Institute of Mathematics and Statistics of the University of São Paulo (IME-USP) and a series of courses (100+) on DataCamp and Coursera. This way, I contact with data science, both its formal base and its applied side. My experience in this field mainly was by using the programming language Python and, to a lesser extent, SQL and R. The following are some activities related to data science that I am very familiar with (they are separated only for clarity): pre-processing (collecting, importing, cleaning and manipulating), exploratory analysis and application of statistical tooling, attribute engineering and modeling (with classical statistics and/or machine learning), and visualization and interpretation of the results. In the machine learning part, I have experience in both supervised and unsupervised learning, from simple tree-based models to different ensemble techniques, as well as neural networks and deep learning. Using Python, I am very familiar with the libraries: numpy, pandas, statsmodels, scipy, sklearn, xgboost, keras, pyspark, matplotlib and seaborn. In addition to many other libraries in different degrees of depth.