Things I build and contribute to

What it is. Geomstats is a Python package for computations, statistics, machine learning and deep learning on manifolds. It splits into a geometry module (manifolds, Lie groups, fiber bundles, shape spaces and Riemannian metrics) and a learning module (statistics and learning algorithms for data that lives on curved spaces), running on NumPy, Autograd or PyTorch backends.

What I did. I joined as one of the early core contributors (3rd contributor to the project) and focused on bringing geometric structure to real data and models:

• Implemented and validated data representation on manifolds, mainly graph-structured data.
• Developed learning methods for complex-system models, including simple, timed and hybrid automata, cast on geometric structures.
• Contributed to the core geometry/learning APIs and their test coverage across backends.

What it is. NeuroWaves provides the documentation, experiment code and processing pipelines for the magnetoencephalography (MEG) and electroencephalography (EEG) users of the NYUAD BioMedical Imaging Core. It standardises how brain-imaging data is acquired, validated and analysed across the lab.

What I did. As Research Scientist I own much of the lab's technical infrastructure and reproducibility:

• Build and maintain reproducible MEG/EEG processing pipelines from raw signals to analysis-ready data.
• Author the lab documentation on Read the Docs, covering acquisition, preprocessing and analysis workflows.
• Enforce BIDS-compliant datasets with automated validation, and set up continuous testing for the NYUAD pipelines.
• Design experimental paradigms and keep the MEG, EEG and OPM systems calibrated and research-ready.