Custom Generative AI Platform

Virginia Polytechnic Institute and State University (Virginia Tech) is soliciting proposals for the design, deployment, and support of a custom generative AI platform leveraging its existing Microsoft Azure cloud infrastructure. This platform will be accessible to eligible faculty, staff, and students, integrating private instances of leading large language models (LLMs) from providers such as OpenAI, Anthropic, and Google within Virginia Tech’s Azure tenant and on-premises HPC environment. The focus is on creating a secure, self-service solution featuring a chat interface, REST API endpoints, and configurable Role-Based Access Control (RBAC) aligned with Virginia Tech’s identity management through Grouper and Microsoft Entra ID. Key platform capabilities include integration with institutional data sources such as Microsoft 365 and Google Drive, enforcement of data retention policies, compliance with accessibility standards (WCAG 2.1/2.2 AA), and strict data residency controls confining all data within U.S.-based cloud regions under Virginia Tech’s custody. The solicitation, posted in December 2025 with proposals due by mid-December 2025 and expected award in early 2026, calls for a turnkey deployment within seven to nine weeks post-award and full system availability by July 15, 2026. The vendor is responsible for the platform design, solution implementation, and API provisioning, while Virginia Tech manages the infrastructure and operational support. The contract outlines stringent security and governance requirements: complete logging and auditability of all user interactions and data access, prohibition on using university data for AI model training, content guardrails, and collaborative security management. Service Level Agreements specify tiered response and resolution times for support incidents. While no formal Federal Acquisition Regulation (FAR) clauses or evaluation factors are detailed, the solicitation encourages proposal submissions from SWaM-certified vendors, requires detailed pricing inclusive of Azure consumption and third-party licenses, and emphasizes compliance with institutional policies on data control and accessibility. The contract explicitly excludes physical packaging, inspection, or shipping elements, focusing exclusively on a digital cloud platform delivered within Virginia Tech’s environment.