Scientific machine learning for electronic structure and atomistic simulation. Combining deep learning, physics-based modelling and high-performance computing to accelerate materials research.
Tools for scientific ML, electronic structure and atomistic modelling.
Machine learning for electronic structure.
An open-source Python framework predicting electron charge and kinetic energy densities from crystal geometry to accelerate electronic-structure calculations.
Everything you need for fermionic quantum simulation.
An open-source Python framework for electronic structure using classical and quantum computing methods, generating data for ML pipelines.
Chemical and geometric descriptors for ML.
An open-source toolkit generating chemical and geometric descriptors for machine learning models of materials.
Visual atomistic modeling.
Builds, edits and inspects atomic structures visually, with a C++ core that stays responsive on systems large enough to stall a typical viewer.
Deep learning, scientific software, computational physics and teaching.
Scientific ML & Deep Learning
Fourier neural operators and symbolic regression for electron charge and kinetic energy densities; graph neural networks for total energies and interatomic forces. End-to-end ML pipelines connect simulation data, model training and physics-based validation on computing clusters.
Python/C++ · Frontend and Backend
Open-source frameworks for electronic structure and materials ML, with applications spanning the numerical core, APIs, interfaces and deployment. Scientific workflows connect data generation, feature engineering and simulation.
Electronic structure of low-dimensional materials, from topological insulators to phosphorene, studied from first principles.
Nine years of higher education teaching across quantum mechanics, thermodynamics, solid state physics and programming.
I am a machine learning engineer and computational scientist with a PhD in Physics and over ten years of international R&D experience in Brazil, the United States and Singapore. At Ilum – School of Science, CNPEM, I work at the intersection of AI for Science and electronic structure.
My work combines Fourier neural operators and symbolic regression for electronic densities with graph neural networks and machine-learning interatomic potentials for atomistic simulation. I build data and ML pipelines in Python and PyTorch, using high-performance computing to scale training and simulation workflows.
I develop open-source scientific software, including Poraquê, Mandacaru and Quasigraph, and build applications from the numerical core through the API, interface and deployment. My experience also includes leading multidisciplinary research teams and teaching physics, Python and C++ at undergraduate and graduate level.
Spontaneous chemical functionalization via coordination of Au single atoms on monolayer MoS2
Sci. Adv. 6, eabc9308
Multiferroic Two-Dimensional Materials
Phys. Rev. Lett. 116, 206803
Vertical twinning of the Dirac cone at the interface between topological insulators and semiconductors
Nat. Commun. 6, 7630