Custom Generative AI Platform

Background
The Division of Information Technology at Virginia Tech aims to deploy a generative AI platform for eligible faculty, students, and staff. This initiative seeks to leverage the existing partnership with Microsoft and the investment in the Microsoft Azure cloud platform to create a generative AI platform that provides access to private instances of frontier large language models (LLMs) from providers such as OpenAI, Anthropic, and Google. The goal is to enhance educational and research capabilities at Virginia Tech by offering advanced AI tools.

Work Details
The contract involves developing a custom generative AI platform that includes:

- A chat-centric web-based user interface for users.
- Direct API access for researchers and developers.
- Access to LLMs trained or augmented using university-managed high-performance computing infrastructure.

Mandatory capabilities include:

- Operation within the existing Microsoft Azure tenant.
- Compliance with Azure AI Foundry best practices.
- Access to product source code for development and deployment of extensions.
- Self-service retrieval augmented generation (RAG) support, allowing users to connect various data sources like Microsoft OneDrive and Google Drive.
- Image and video generation capabilities based on model selection.

Monitoring requirements include:

- Near-real-time reporting on model usage and user queries.

Security measures such as single sign-on through Microsoft Entra ID and role-based access control are required. Ongoing technical support and maintenance are also part of the contract obligations.

Period of Performance
The contract will be performed over a period of 2 years, with options for four additional one-year renewals based on negotiations.

Place of Performance
The services will be performed at Virginia Tech's facilities, utilizing the Microsoft Azure cloud infrastructure.

Bidder Requirements
Bidders must register as vendors within the eVA system prior to proposal submission. They are encouraged to utilize SWaM subcontractors if applicable, maintain their SWaM certification throughout the contract duration, and provide insurance coverage including general liability and workers' compensation as specified in the solicitation.