Bioinformatics engineer · Biomedical data science · Scientific Python & R

I am a graduate of École d’Ingénieur Denis Diderot, Université Paris Cité, specialising in bioinformatics. I work at the intersection of biological data, quantitative methods and scientific software.

My research experience includes proteomic and clinical biomarker modelling at Brain&Mind, quantitative muscle MRI at Institut de Myologie, and genomics at Paris Brain Institute. The public projects below are separate portfolio studies: they document methods, contributions, results and reproducibility.

A Snakemake workflow comparing matched Illumina and Nanopore yeast reads, from QC and alignment to filtered SNP comparison and targeted discordance review.

• Evidence: recorded real-data results, source fingerprints, unit tests and synthetic integration CI.
• Finding: comparison eligibility changes markedly with ONT base-quality thresholds; filtering explains selected discordances.
• Tools: Python, Snakemake, BWA, minimap2, samtools, bcftools.
• Scope: exploratory ONT calling without an independent truth set.

A public-data modelling study with training-only preprocessing, nested group-aware validation, a dummy baseline and four sensitivity analyses.

• Evidence: out-of-fold predictions, tests, reproducible figures and a documented complete refit.
• Finding: modest primary discrimination and substantial inflation when current diagnostic variables are included. Numerical refit differences are documented.
• Tools: Python, scikit-learn, XGBoost, pandas, uv.
• Scope: internal methodological evaluation; no external clinical validation.

A CLI joining PROSITE, UniProtKB, PDB and SIFTS to inspect domain coverage and structural annotations.

• Evidence: dated source fingerprints, offline replay and tests under Python 3.10 and 3.12.
• Finding: an empty primary PROSITE PDB field does not mean structures are absent; identifier overlap needs residue-range checks.
• Tools: Python, Biopython, biological APIs.
• Scope: annotation and coordinate analysis, without docking or affinity prediction.

For genomics and R roles, start with Yeast WGS and Viral Genome Comparator. For Python development, the assembler provides complementary algorithmic evidence.

Demonstrated in these repositories: scientific Python and R, biological data processing, graph algorithms, predictive-model evaluation, workflow automation, testing, GitHub Actions, dependency management and execution provenance.

Broader training and experience: SQL, biomedical statistics, quantitative imaging and data platforms. I am also developing a data-engineering direction; the repositories above primarily demonstrate scientific computing and biological analysis.

I am interested in bioinformatics, biomedical data science, scientific software, research engineering and relevant R&D opportunities.

LinkedIn · Email