IT Services for AI Benchmarking Competition

This contract opportunity involves providing support through a cloud platform for conducting an AI benchmarking competition focused on the application of Foundation Models in the security sector. Foundation Models are complex neural networks trained on large datasets, characterized by their generality and versatility. These models contain hundreds of billions of parameters that represent implicit knowledge acquired from the data, serving as the foundation for further applications. The overall process will last up to 36 months, with decisions on its continuation made in the 9th and 20th months of the competition. The start of this support service is linked to the commencement of the overarching research program, currently planned for the fourth quarter of 2025. The goal of this overarching process is to deliver innovative research and development services in a competitive framework for the comprehensive evaluation of generative Foundation Models in a security context. This tender specifically seeks to provide the central benchmarking platform for conducting the challenge.