Research areas

  • Materials discovery using first principles simulations and surrogate models (MODNet, re2fractive).
  • Materials and chemical data management for experiment and theory (datalab, datatractor, MADICES).
  • Crystal structure databases and associated software development for reproducible science (OPTIMADE).
  • High-throughput computational materials science workflows and crystal structure prediction for battery materials with AIRSS, genetic algorithms and other related methods (PhD thesis).

Software

Most of my software work can be found on my GitHub.

  • datalab is a web API and Vue front-end app for materials chemistry data management, tracking samples, devices and their connections, and performing simple analysis of characterisation experiments (XRD, NMR, electrochemistry). It is currently deployed for 6 different groups worldwide.
  • jobflow-remote is a distributed high-throughput workflow manager built on top of the Jobflow library, enabling remote job submission and orchestration.
  • optimade-python-tools is a Python library of tools for implementing and consuming OPTIMADE APIs, which I develop and maintain.
  • OPTIMADE is the specification repository for OPTIMADE, a common REST API for access to materials databases, which I help to develop and maintain. [7]
  • optimade-maker is a set of tools for converting static archived data into OPTIMADE APIs; this tool is used in production at the Materials Cloud Archive and was the result of a reserach visit to the Paul Scherer Institut.
  • odbx.science is a public OPTIMADE API that hosts the Morris group’s crystal structures and an increasing number of materials discovery datasets [misc], [GNoME] and thismaterialdoesnotexist.com.
  • datatractor / MaRDA Extractors WG: I founded and co-lead a MaRDA working group (open to all!) on the topic of metadata extractors for materials science and chemistry, see the schemas and registry.
  • MODNet is a package for materials property prediction focused on small datasets, which I help develop and maintain. [6]
  • matador is a Python package for the aggregation, standardisation and analysis of the results of first-principles computations, with a focus on battery materials [5].
  • ilustrado is a Python package that implements a highly-customisable, massively-parallel genetic algorithm for ab initio crystal structure prediction.

I additionally help to maintain several other open source projects, including galvani (parsing battery file formats in Python), matminer (materials machine learning in Python), PASCal (fitting of in situ diffraction strain data [11], as well as making frequent contributions to upstream packages in the field.