Securely running AI algorithms for 100,000 users on private data

Radboud University Medical Center Nijmegen (Radboudumc) developed a secure AI platform called grand-challenge.org that enables researchers to run and validate AI algorithms on sensitive medical data while maintaining strict privacy and security standards.

• Platform serves 100,000+ global researchers and developers
• Allows secure testing of AI algorithms on private medical datasets
• Uses AWS services like Amazon SageMaker, EventBridge, and ECS for infrastructure
• Hosted 130+ public challenges with 10,000+ algorithm submissions
• Processed 45.6 terabytes of protected medical data
• Enables algorithm testing without exposing intellectual property

The solution demonstrates how cloud technologies can facilitate secure, scalable AI research in sensitive domains like healthcare, while maintaining data privacy and compliance.