Junior Machine Learning Engineer
Research & Development for the implementation of a hyperlocal air quality modeling via an innovative supervised hybrid approach.---Realization of a functional prototype quite more efficient than its predecessor, focusing on the prediction of ozone (O3), nitrogen dioxide (NO2) and fine particles (PM10 and PM2.5). Innovative hybridization merging the use of Particle Swarm Optimization (PSO) bio-algorithms and neural networks (NN) implementing the concept of calculation by reservoir; all adapted to the use of geographically projected data. This work was mainly done in Python with Pandas, Numpy, Rasterio and Scikit-Learn for main libraries, although the Machine Learning algorithms implemented were for the biggest part coded from zero, for a better match to the specificities of the project.