The Salk Institute embraced VAST Data to modernize and future-proof its AI and big data research infrastructure. Faced with growing data needs and limitations of legacy systems, VAST's Disaggregated Shared-Everything (DASE) architecture provided scalable, high-throughput, low-latency storage optimized for AI/ML workloads. Talmo Pereira, a Salk Fellow and Principal Investigator, highlighted tangible improvements in AI/ML pipelines, including faster data streaming from centralized storage to GPU nodes, which accelerates model development. VAST’s platform simplifies data management with a unified global namespace supporting seamless data sharing across protocols and collaborators.

Join for free to read