At Imagem Geosistemas, I designed, trained, and deployed semantic segmentation models for high-resolution imagery covering agricultural areas, urban features, and environmental change. I also translate business requirements into spatial analysis and AI solutions.
At Media.Monks, I built Marketing Mix Models and attribution models to guide budget allocation across Google and Meta Ads. I also maintained decision support models, automated dashboards, and trained data analysts on statistical modelling.
As a PhD researcher at Computing Department, UFSCar, I developed curvature-based methods for manifold learning and message aggregation in heterophilic GNNs. My research includes SH-UMAP and ACN-GCN, with publications at SIBGRAPI.
