Recruitment of a Training Provider for NVIDIA AI Infrastructure & Operations

Recruitment of a Training Provider for NVIDIA AI Infrastructure & Operations Request for proposal Description RECRUITMENT OF A TRAINING SERVICES PROVIDER FOR THE NVIDIA AI INFRASTRUCTURE AND OPERATIONS: DEPLOYMENT, VIRTUALIZATION, AND MANAGEMENT TRAINING Item Description Office / Unit / Project UNDP Rwanda Place Venue to be proposed by the training services provider, meeting the minimum technical and venue requirements Type of Contract Local/International Consultancy Firm Expected Duration Up to 5 months from contract signature 1. Background The UNDP, through the Timbuktoo framework, is investing in sovereign Artificial Intelligence infrastructure to expand the computational capacity of African countries and accelerate progress toward the Sustainable Development Goals. Standardized, high-performance infrastructure is being deployed in six partner countries: Rwanda, Kenya, Malawi, Zambia, Togo, and South Africa enabling local developers, researchers, and innovators to test and deploy AI solutions in priority sectors such as agriculture, climate action, ensuring localized data processing, secure storage, and technological sovereignty. In 2026, an inaugural AI infrastructure workshop was convened, bringing together technical counterparts from all six partner countries. That workshop surfaced a clear need: national technical teams require deeper, hands-on operational skills to independently deploy, virtualize, monitor, and sustain the newly installed AI/HPC infrastructure. 2. Rationale Sovereign AI infrastructure delivers value only when local teams can operate it independently and reliably. Infrastructure, platform, and operations personnel in partner countries currently have varying levels of exposure to enterprise-grade GPU infrastructure management. Without structured, hands-on training, countries risk: • Prolonged dependence on external vendors and implementation partners for routine deployment, configuration, and troubleshooting; • Underutilization of installed HPC/GPU capacity; • Inconsistent operational and security standards across the six national data centers; • Recurring, avoidable operational issues for example, GPU topology misconfiguration, already reported across more than one country office going unaddressed for lack of in-house diagnostic skill. Therefore, there is a need for a standardized, instructor-led training. The assignment is a customized, hands-on technical training service. The successful provider shall demonstrate the technical ca