Douglas David Baptista de Souza

Introduction

About, Dr. Baptista de Souza is a PhD Engineer with R&D experience in Machine Learning and Signal Processing. Having participated in and led many R&D projects in the academia and the industry, Dr. Baptista de Souza is author and reviewer of papers in world-class conferences and journals (IEEE, Elsevier), and international consultant in Machine Learning for several companies. Currently, Dr. Baptista de Souza is a Senior Machine Learning Scientist at Dynamox, Brazil. Professional experience - Industry

2020 Senior Machine Learning Scientist, Dynamox, Brazil. 2018–2020 Lead Machine Learning Engineer, GE Renewable Energy, Brazil. Responsible for developing machine learning solutions (e.g. anomaly detectors, fault forecasters) and statistical analyses (e.g., component life models) for wind turbines and the digital services team. 2015–2017 Researcher, GE Global Research Center, Brazil, Worked on machine learning and signal processing R&D, such as tracking filters for system identification and predictive/diagnostic models.

{ Led a project to create a fault forecasting model for wind turbines belonging to the GE fleet in LATAM.

{ Worked on the development of a digital twin for one of the most advanced GE wind turbines, in a global collaboration with different teams of researchers, engineers and developers.

{ Developed scripts to classify aircraft final approaches, implementing the developed analytic on a platform with 60 TB of flight data, used by major airline companies around the world.

{ Worked on the development a digital twin for a GE jet engine, one of the first to be developed in LATAM.

{ Used machine learning models to predict post-flight customer satisfaction for a major airline company, yielding important insights on customer profile and flight experience.

{ Developed an AI model for automatic detection of anomalous lubrication conditions in gas turbines, which allowed more assertive and faster interventions by remote monitoring teams. Professional experience - Academia

2015–2015 Assistant Professor of Physics, Federal University of Santa Catarina, Florianópolis, Brazil.

{ Waves and propagation processes, classical mechanics. 2015–2015 Professor of Electrical Engineering, SATC University, Criciúma, Brazil.

{ Electromagnetic waves and design of antennas.

International Consultant

Professional

2019 AI/NLP specialist, Company: Swae, Canada.

{ Supported the development of an interactive NLP pipeline with diverse functions. 2018-2019 Machine Learning & Speech Signal Processing Specialist, Company: Oyalabs, Hong Kong.

{ Developed a framework for automatic detection of silent/speech segments in audio files. 2018 Machine Learning & Signal Processing Specialist, Company: Inspirit IOT, United States.

{ Studied and worked on the development of techniques to identify and process rare events in audio signals. 2014 Machine Learning Specialist, Chaordic company, Florianopolis - Brazil.

{ Carried out a study to find structural patterns and change points in online purchase and click time serise. Post-doctoral fellowships

2013–2014 LTHE lab, Post-doctoral fellow, Université Joseph Fourier, Grenoble, France. Research topic: Developed a Matlab toolbox for smart selection of signals simulated from a particular run of a Global Climate Model (GCM), considering many aspects like emission scenarios and downscale rules. The framework was a multi-stage cluster selection method developed during the post-doctoral assignment. Rua Santos Saraiva 238 ap 502, Florianópolis – SC, Brasil B adcr46@r.postjobfree.com - www.linkedin.com/in/douglasdbs - +55-48-996****** Education

2010–2013 PhD in Electrical Engineering (Signal Processing), Institut National Polytechnique de Grenoble, GIPSA-Lab, Grenoble, France.

{ Thesis: On nonparametric techniques for analyzing nonstationary signals. Advisor: Jocelyn Chanussot

{ Research area: Statistical Signal Processing, Machine Learning, Real-world Signal Analysis. 2005–2009 Electrical Engineering degree, Federal University of Santa Catarina, Brazil. Certifications

2016 Six Sigma Green Belt (DFSS and Lean), General Electric (GE). Awards

2010 Erasmus Mundus, Erasmus Mundus full PhD scolarship in Grenoble, France. Languages

Fluent Portuguese (Native language), English, and French Computer skills

Languages: Python, MATLAB, R Other

skills/packages

Tensorflow, Keras, LaTeX

Publications (all as first author)

Journal

paper

IEEE Signal Processing Letters, An improved stationarity test based on surrogates, D. B. de Souza, J. Chanussot, A.-C. Favre, and P. Borgnat, vol. 26, no. 10, p.1431-1435, 2019. Journal

paper

IEEE Signal Processing Letters, A time-varying autoregressive model for characterizing non- stationary processes, D. B. de Souza, E. V. Kuhn, and R. Seara, vol. 26, no. 1, p.134-138, 2019.

Journal

paper

Elsevier Signal Processing, A nonparametric test for slowly-varying nonstationarities, D. B. de Souza, J. Chanussot, A.-C. Favre, and P. Borgnat, vol. 143, February 2018, p.241-252. Conference

paper

IEEE ICASSP 2014, A new nonparametric method for testing stationarity based on trend analysis in the time marginal distribution, D. B. de Souza, J. Chanussot, A.-C. Favre, and P. Borgnat. In: 2014 IEEE International Conference on Acoustics, Speech and Signal Processing

(ICASSP), p.320, Florence, Italy.

Conference

paper

IEEE ICASSP 2014, On selecting relevant intrinsic mode functions in empirical mode decompo- sition: An energy-based approach, D. B. de Souza et al. In: 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), p.325, Florence, Italy. Conference

paper

IEEE LASCAS 2014, On generating a finite pulse or a symmetric impulse response by a generalized approximation function, D. B. de Souza, and S. Noceti Filho. In: 2014 IEEE 5th Latin American Symposium on Circuits and Systems (LASCAS), p.1, Santiago, Chile. Conference

paper

IEEE ICASSP 2012, A modified time-frequency method for testing wide-sense stationarity, D. B. de Souza, J. Chanussot, A.-C. Favre, and P. Borgnat. In: 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), p.3409, Kyoto, Japan. Conference

paper

IEEE ISCAS 2011, An optimum linear phase approximation with small delay obtained by the manipulation of all-pass Padé approximants, D. B. de Souza and S. Noceti Filho. In: 2011 IEEE International Symposium on Circuits and Systems (ISCAS), p.2265, Rio de Janeiro, Brazil. Conference

paper

SBRT 2009, Forma Simplificada de Determinacao das Funcoes de Atraso Filanovsky-Matkhanov e Propostas de Modificacoes, D. B. de Souza and S. Noceti Filho. In: XXVII Simposio Brasileiro de Telecomunicacoes (SBRT), p.1-6, Blumenau, Brazil. Patents

2018 General Electric Co, System and method for detecting lubricated bearing condition, Subrat Nanda, Douglas David Baptista de Souza, Bruno Paes Leao, 2019. https://patents.google.com/patent/US20180334917A1/en Rua Santos Saraiva 238 ap 502, Florianópolis – SC, Brasil B adcr46@r.postjobfree.com - www.linkedin.com/in/douglasdbs - +55-48-996******

Contact this candidate