Projects
Research Projects
Problem. Predicting agricultural harvest outcomes and supporting related field decisions typically relies on manual observation and historical rules of thumb, with limited use of computational forecasting.
Context. Research project, Universidade de Pernambuco (UPE), 2005–2006.
My role. Undergraduate research (iniciação científica) student for two years, supervised by Prof. Fernando Buarque de Lima Neto.
Methods. Neural networks applied to agricultural time series for crop-harvest prediction and decision support; a mobile data-visualization component using georeferenced GPS data.
Outputs. An intelligent decision-support prototype for crop-harvest prediction.
Selected Applied Projects
Problem. Vendors and public bodies offering software through the Brazilian Public Software program (Software Público Brasileiro) needed clear, up-to-date guidance on the process and requirements for listing software on the platform.
Context. Brazilian Ministry of Planning, Software Público Brasileiro program, 2016.
My role. Member of the Software Público Brasileiro team; technical lead for the version 3.0 update of the Manual do Ofertante (Vendor’s Manual).
Methods. Technical writing and process documentation, in coordination with the program team.
Outputs. Manual do Ofertante, v3.0 (PDF).
Additional applied work at Ipea (data integration, data quality, pipeline automation) is not yet described here, pending a public, non-operational framing.