Medical Imaging AI Services

The vendor is required to provide medical imaging AI solution and associated software for training and deploying very large AI foundation models.
- Medical imaging research, the System needs to support all common open-source AI software tools.
- System must meet the following technical specifications:
•    Be equipped with eight (8) high-performance NVIDIA b300 graphics processing units (GPUS), optimized for AI training and inference tasks.
•    Each GPU must include a minimum of 288 GB of high-bandwidth memory, to support the training of very large AI models.
•    The system must be equipped with high-performance central processing units (CPUS) to support fast data loading and efficient parallel computing.
•    The ideal configuration includes at least two (2) intel Xeon platinum 6776 processors or higher-performance equivalents.
•    To accommodate large-scale medical imaging datasets used in AI training, the system must include a minimum of 30 tb of high-performance internal storage.
•    The system must support the latest open-source software libraries for training AI foundation models, including but not limited to PyTorch and TensorFlow.
•    Total power consumption must not exceed 14.3 kw.
•    The system must include a comprehensive on-site warranty of no less than three (3) years, covering – at a minimum - parts, labor, travel (if applicable), and full support for both hardware and software components. 
•    All associated costs must be fully included in the warranty with no additional charges to university.

- Questions/Inquires Deadline: September 25, 2025