Cheminformatics Scientist (PhD) · Reaction informatics · Generative GNNs · Synthesis planning · Data/MLOps for discovery
GitHub · Google Scholar · LinkedIn · Email
I build production-grade cheminformatics tooling and AI workflows for small-molecule discovery — from raw reaction/assay data to reproducible ML pipelines and deployed services.
- Now: Cheminformatics Data Engineer @ Qubit Pharmaceuticals (Paris) — secure data platforms, shared schemas, MLOps (MLflow), CI/CD.
- Before: Postdoc with Eli Lilly × Laboratory of Chemoinformatics — ultra-large chemical space navigation (GTM), GNNs for DEL, AI-designed antimicrobial peptides.
- PhD (Univ. of Strasbourg): Generative GNNs + retrosynthesis — design molecules and plan how to make them.
- SynPlanner — end-to-end synthesis planning stack: reaction curation → rule extraction → model training → GNN-guided MCTS planning.
- HyFactor — graph-based generative model for molecular structure generation.
- VQGAE — vector-quantized graph autoencoder for learning fragment spaces for molecule design.
- CGRtools / CIMtools — reaction & QSPR tooling contributions (data curation, validation, bugfixing).
- ChemographyKit - implementation of GTM on Pytorch for chemography applications