ALMEIDA DE MENDONCA Joao Paulo

Research activities

I work in computational physics, applying it to atomistic modeling in materials science. Most of my time is spent coding and helping other researchers improve their data-management tools and machine-learning applications.

An important part of my work is the use of bio-inspired algorithms (mostly evolutionary or particle-swarm algorithms) and machine-learning approaches (e.g., ANNs, clustering, and manifold learning), combining them with well-known classical and first-principles calculation software (both packages like LAMMPS, VASP, xTB, and ORCA, and libraries like PySCF).

Recently, I have been doing research on the development of machine-learning functionals to model systems with spin crossover via DFT. I work as a research engineer on the DIAMOND project, developing machine-learning workflows with other materials-science researchers.