Case Study: ZIFF achieves breakthrough AI analytics speed and scalability with Pure Storage FlashBlade

Ziff Breaks Through The Ai Analytics Barrier With Flashblade From Pure Storage

ZIFF is an AI and deep‑learning services firm that helps organizations lacking in‑house analytic horsepower turn massive image, video and audio data sets into actionable insights. The company struggled with persistent storage I/O bottlenecks and cumbersome data shuffling that kept GPUs and data scientists idle, slowed projects to weeks, and prevented clients from using AI at scale.

By adopting Pure Storage FlashBlade as its core data platform, ZIFF eliminated the I/O constraint and dramatically accelerated processing: a 7‑blade FlashBlade (70+ TB usable) enabled indexing 4 million images in 24 hours and cleaning a 14 million‑image training set in under a day. Throughput of ~3,000–4,000 images/sec has tripled ZIFF’s workload capacity, cut model training from weeks to 24–48 hours, increased data‑scientist productivity, and delivered faster, lower‑cost results for clients.

ZIFF

David Gonzalez

CEO