Jules Samaran
PHD Student at École normale supérieure
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolJules Samaran is a PhD student in Computational Biology at Institut Pasteur and ENS, combining seven years of experience in machine learning, statistics, and single-cell omics. He develops multimodal integration methods under academic supervision while contributing to prominent open-source tools like scvi-tools, where he has fixed preprocessing bugs and improved posterior summaries for single-cell RNA-seq analysis. His background spans research internships and visiting roles—from Criteo and Booking.com to UC Berkeley and Osaka University—applying deep generative models, visual grounding, and anomaly detection across domains. Comfortable bridging theory and practice, he leverages engineering rigor from a Mines ParisTech engineering degree and an M.Sc. in Applied Mathematics, Vision and Learning to make reproducible, production-ready research. An understated strength is his knack for pragmatic fixes in community projects that materially improve downstream analyses.
7 years of coding experience
2 years of employment as a software developer
Msc MVA: Applied Mathematics, Vision and Learning, Msc MVA: Applied Mathematics, Vision and Learning at École normale supérieure Paris-SaclayMathematics, Mathematics at Lycée Pierre De FermatDiplôme d'ingénieur, Mathematics and Computer Science, Diplôme d'ingénieur, Mathematics and Computer Science at MINES ParisTech
English, Spanish, Japanese
Github Skills (10)
rna-seq10
pandas10
seq10
deep-learning10
python10
sc10
numpy10
data-analysis10
scikit-learn9
scikit9
Programming languages (4)
Github contributions (5)
scverse/scvi-tools
Sep 2018 - Jan 2019
Deep probabilistic analysis of single-cell and spatial omics data
Role in this project:
Data Scientist
Data ScientistContributions:60 commits, 23 PRs, 80 pushes in 4 months
Contributions summary:Jules primarily contributed to bug fixes and improvements within the scvi-tools repository, focusing on single-cell RNA-seq data analysis. They resolved issues related to data preprocessing, including unzipping data, handling zero-count cells, and replacing deprecated functions. Their contributions also involved modifications to the model's functionalities, such as saving and outputting posterior summary statistics and correcting batch mixing calculations. They made some changes to the dataset loading process as well.
hierarchicaldeep-generative-modelmixture-of-expertssingle-cell-rna-seqsingle-cell
jules-samaran/VL-BERT
Apr 2020 - Sep 2020
Code for ICLR 2020 paper "VL-BERT: Pre-training of Generic Visual-Linguistic Representations".
Contributions:6 PRs, 109 pushes, 5 branches in 5 months
nlpiclrrepresentationsbertlinguistic
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.