Machine learning and atomistic materials simulations

Welcome

We are a computational research group working at the interface between machine learning and atomistic simulations. We use the tools of data science and engineering as well as physics-based simulations like density functional theory and molecular dynamics to design and understand materials. 

Designing for function.

We use computational tools to tackle design of materials in complex combinatorial search spaces, such as organic electronic materials, energy storage polymers and molecules, and heterogeneous (electro)catalysts. In addition to screening large numbers of possible candidates with forward models (given a material, predict its properties), machine learning tools allow us to address the inverse design question, that is, given the desired properties, imagine the material.

Basic questions.

The combination of large experimental datasets and accurate theoretical simulations with statistical inference allows computers to answer fundamental scientific questions like never before. We ask questions about the nature of chemical reactivity, materials transformations, and ion transport.


Active Learning

Representation Learning

Generative Models and Inverse Design

High-throughput virtual screening


We are recruiting! 

We are actively looking for new group members at all levels.

  • MIT undergraduate students interested in joining the group for a UROP please shoot Rafa an email.
  • Undergraduate students interested in joining the group for their PhD are encouraged to apply for admission to an MIT graduate program.
  • Graduate students admitted to DMSE and also to other departments at MIT who are interested in working with us should contact Rafa. 
  • We are also on the lookout for outstading postdoctoral researchers with a background in any combination of the following: machine learning, high-throughput simulation, materials informatics, electronic structure. Python chops and experience using databases are a big plus. In particular, we searching to fill funded positions in the spaces of 
    • Machine learning, DFT and molecular dynamic simulation of vacancy formation and atomic exchange in refractory high entropy alloys (from September 2022 to August 2024)
    • Molecular dynamics and grand canonical simulations of epitaxial growth of III-V semiconductors with machine learning potentials (from September 2022 to August 2023)