A few of the things I work on, in and around high-performance computing. Most are open source; the rest live inside SHARCNET and Digital Research Alliance of Canada infrastructure.
ViewClust
A Python package for computing and visualizing usage metrics on Slurm-based HPC clusters. Originally built at SHARCNET to standardize how cluster utilization was measured across analyst data frames; now also used by collaborators at WestGrid, Calcul Québec, and MILA. Companion package ViewClust-Vis adds a set of summary figures.
- Source: https://github.com/Andesha/ViewClust.
- Companion: https://github.com/Andesha/ViewClust-Vis.
PyLossless
Python tooling for reproducible EEG preprocessing and quality-control workflows. Aimed at making the pre-analysis steps of EEG studies auditable and easy to re-run as data and pipelines evolve.
EEGStudyFlow
Workflow patterns and tooling for managing EEG studies end-to-end, from raw recordings through preprocessing and quality control to analysis-ready outputs that researchers can hand off without re-deriving state.
SHARCNET Analytics
Internal analytics and operational-insight work supporting research computing systems and the people who use them: usage forecasting, fair-share investigations, scheduler tuning, and ad-hoc deep dives into cluster behaviour.