Hamza Cherkaoui
Postdoctoral Researcher at Télécom SudParis
Paris, Ile-de-France
Summary
👤
Senior🎓
Top SchoolHamza Cherkaoui is a Paris-based postdoctoral researcher with nine years of experience at the intersection of machine learning, statistics, and signal processing, currently investigating collaboration in dynamic settings, diffusion models and curriculum learning at Télécom SudParis. He holds a PhD from CEA’s Parietal team where he developed TV-regularized semi-blind deconvolution methods for whole-brain fMRI analysis, and has extended that expertise to bandits, Bayesian optimization and LLMs during postdoctoral stints at Huawei Paris and Université Paris-Saclay. Hamza blends theoretical work on Markov chains and bandit problems with practical software contributions—such as examples and fixes to the widely used nilearn neuroimaging library—so his research is grounded in reproducible tooling. He combines a strong background in applied mathematics and optimization with hands-on engineering from internships and projects (multi-threaded tree growth, MRI reconstruction, embedded systems), enabling him to translate complex models into usable code and experiments.
9 years of coding experience
5 years of employment as a software developer
Master 2 (M2) Data Sciences, Master 2 (M2) Data Sciences at École PolytechniqueMaster 2 (M2) Applied Mathematiques, Master 2 (M2) Applied Mathematiques at University of Lille 1 Sciences and TechnologyCentrale Lille
English, Arabic
Github Skills (10)
machine-learning10
python10
neuroimaging10
mri10
brain-imaging9
scikit-learn9
scikit9
glm8
connectivity8
numpy7
Programming languages (2)
Github contributions (5)
nilearn/nilearn
Jan 2018 - Jan 2018
Machine learning for NeuroImaging in Python
Role in this project:
Data Scientist
Data ScientistContributions:5 commits, 2 comments in 6 days
Contributions summary:Hamza contributed to the nilearn repository by implementing and refining examples related to machine learning for neuroimaging. Their work included adding an example for extracting the Default Mode Network (DMN) from the ADHD dataset, demonstrating the application of GLMs and contrast estimation. Additionally, the user addressed issues in the nistats library by modifying design matrix related documentation and fixing paradigm handling, likely to improve the usability and accuracy of statistical analyses. Furthermore, the user improved code clarity and added warning messages.
neuroimagingpythondecodingbrain-connectivitymachine-learning
hcherkaoui/hcherkaoui.github.io
May 2020 - Sep 2023
Contributions:161 pushes, 7 branches in 3 years 5 months
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.