Materials Intelligence Research

The Materials Intelligence Research (MIR) group at Harvard University develops and uses computational methods combining quantum physics & chemistry and statistical mechanics with machine learning to understand and design complex materials.

Our goal is to create efficient algorithms and accurate atomistic models to be able to describe systems of surfaces, liquids, solids, and molecules for applications in energy storage and conversion.

Open positions

Deep equivariant neural networks, Bayesian force fields, active learning, exascale reactive MD

Solid-state, ionic liquid, polymer electrolytes, correlated transport

Dynamics of nanoparticles, surface restructuring, heterogeneous reactions, 2D materials

Electrical conductivity, thermoelectric and magneto-transport properties.

Machine learning for development of fast and accurate density functionals

Open-source computational tools and frameworks