In addition to identifying and validating business problems that can be solved with Machine Learning or Deep Learning, I am responsible for coordinating and providing support to teams of data scientists assisting in the implementation of product repositioning and assortment projects, goal building, KPI and evaluation metrics.Among the technologies, frameworks, and methodologies used are:- Databricks;- Spark;- Scikit-learn;- spaCy- Tensorflow;- Keras;- PyTorch;- Python;- SQL;- Agile development.

Synopsis helps you find new reading.synopsis-ia.com

The scientific career at the University of Brasília goes through several research applications, but all converge on using images/videos, AI, and modeling.Master in Mechatronic Systems, graduated in Computer Engineering from the University of Brasília (UnB), I have been a researcher at the Laboratory of Images, Signals, and Audio (LISA) of the Department of Computer Science - UnB since 2014. Among the works carried out are:- Unsupervised classification of events in video surveillance images;- Recognition of actions in videos;- Emotions classification;- Extensive analysis of points of interest in space-time;- Document classification;- Estimation of steering wheel angle of autonomous vehicles;- Classification of diseases in plantations.In addition, I worked as a researcher and developer of Deep Learning models at the Laboratory of Spatial Information Systems (LSIE) of the Department of Geography - UnB, performing satellite image analysis and classifying urbanization regions.Ph.D. in Mechatronic Systems from the University of Brasília, my research focus was on architectures of convolutional neural networks with multiple inputs and the way in which the signal from these inputs is propagated within these networks. My last academic contribution includes the proposition of an internal signal transformation that provides faster convergence, stability, and robustness of the network.Among the technologies and tools used are:- OpenCV;- MATLAB;- Scikit-learn;- Tensorflow;- Keras;- C++;- Python.

Allocated in the Artificial Intelligence Business (NIA) area at Banco do Brasil, I develop Artificial Intelligence models for image and text applications. The work also includes the treatment of these input data, involving Image Processing and Computer Vision techniques, in addition to Natural Language Processing.AI products are used in internal bank processes and have many different intents: from classification and similarity between documents to optical character recognition (OCR) and named entity recognition (NER).Among the technologies, frameworks and methodologies used are:- Tensorflow;- Keras;- Darknet;- PyTorch;- FastAI;- OpenCV;- Scikit-learn;- Python FastAPI;- Agile development.

O projeto ParkCode envolveu o uso de técnicas de Visão Computacional para vídeos e a criação de modelos especialistas em detecção de objetos e reconhecimento ótico de caracteres (OCR). A ferramenta final foi transformada posteriormente em produto comercial.Entre as tecnologias e frameworks utilizados, estão:- Tensorflow- Keras- Python- OpenCV- Scikit-learn

Utilização da ferramenta ORACLE APEX para a criação de sistemas de apoio aos processos internos do TCU.

Desenho de interfaces gráficas usando GIMP e Inkscape; Diagramação em XHTML + CSS nos padrões W3C e de Acessibilidade; Programação em Javascript/Ajax com JQuery.