Pioneering AI-Powered Cancer Diagnostics with VAST Data

Lunit consolidated data, eliminated random read bottlenecks, and maximized GPU performance for AI diagnostics with VAST

Industry

Healthcare & Life Sciences

Use Case

Artificial Intelligence
VAST Data image

Accelerating Medical AI Research

Overview

Lunit, a medical AI pioneer founded in 2013, delivers deep learning solutions like Lunit INSIGHT and Lunit SCOPE to revolutionize cancer diagnosis and treatment. Operating across 40+ countries, Lunit’s high-precision AI models rely on efficient processing of complex medical imaging datasets.

Solution

Lunit adopted the VAST Data Platform to replace unstable, legacy storage that bottlenecked GPU cluster performance. VAST provided a resilient, single-cluster data architecture in South Korea, connecting 1.3 PB of capacity directly to 56 NVIDIA deep learning servers.

Results

VAST eliminated slow random read bottlenecks, allowing Lunit to maximize GPU utilization across its H100 and A100 server clusters. The platform achieved a 2.5:1 data reduction ratio, reducing storage requirements by 60%, while zero-trust security simplified ongoing management.