Four open-source research repositories released with my papers. They contain full PyTorch setups for generative ML models trained on molecular data, workflows for genetic algorithms and quantum-chemistry calculation pipelines.
- tmcinvdesDual-objective inverse design of transition metal complexes: training-set creation, complex assembly, ORCA DFT labelling and analysis (JACS Au 2025).
- JT-VAE-tmcinvdesPyTorch junction-tree VAE, unconditional and conditional, that generates ligands optimised for target properties.
- xyz2mol_tmStructure-to-SMILES conversion for metal complexes (Hückel, NBO and CSD-fix approaches), with SMILES datasets and ML benchmarks.
- genetic_algorithm_for_nitrogen_fixationGenetic algorithm with xTB scoring on SLURM that discovers Mo-based catalysts for nitrogen fixation (Chemical Science 2024).
Python
PyTorch
Scikit-learnSciPy
NumPy
PandasWandBMLflowMolecular EmbeddingsQSPR/QSARRDKitORCAxTBSLURM
