Valerie is PhD candidate in the Software Engineering at Concordia University. Her work focuses on creating open source software tools to adapt High Performance Computing clusters to Big Data neuroimaging pipelines. To achieve this goal, Valerie intends to build a filesystem and scheduler that will minimize the cost of data transfers in neuroimaging pipelines.
Neuroimaging data processed on scientific computing clusters is typically stored on high-performance network-based file systems. Although high performance, these filesystems can be a significant bottleneck when processing large amounts of neuroimaging data. As a CONP scholar, I aim to improve the execution time of Big Data neuroimaging pipelines, through the design and implementation of an open source filesystem and scheduler that will minimize access to high-performance filesystems by prioritizing local storage such as memory and disks.