Yes, there are Software Development Kits (SDKs) and Application Programming Interfaces (APIs) available for development on NVIDIA’s Vera Rubin platform. The platform is designed as a comprehensive full-stack AI supercomputing solution, and as such, it integrates various software components, including libraries and APIs, to facilitate development and deployment of agentic AI workloads. This extensive software ecosystem is crucial for developers to leverage the platform’s advanced capabilities, which span across its seven breakthrough chips, including the Vera CPU and Rubin GPU, and its rack-scale systems.
Key examples of development tools include the NVIDIA DOCA (Data Center Infrastructure on a Chip Architecture) software framework and SDK. DOCA enables developers to build, deploy, and accelerate secure, software-defined data center services using open APIs and hardware offloads, specifically targeting components like BlueField DPUs and ConnectX devices that are integral to the Vera Rubin platform. Furthermore, NVIDIA’s broader strategy for AI development, as evidenced by resources like build.nvidia.com, provides API access for running inference on advanced AI models and offers blueprints with code samples for building AI applications. This indicates a commitment to providing programmatic interfaces for interacting with NVIDIA’s AI infrastructure.
The Vera Rubin platform’s emphasis on agentic AI workflows also drives the need for robust API and SDK support. For instance, Microsoft, a key partner deploying Vera Rubin systems, offers “Voice Live API integration with Foundry Agent Service” which allows developers to create voice-first, multimodal, real-time agentic experiences within environments supported by Vera Rubin. Similarly, system manufacturers like ASUS, who are building AI infrastructure around the Vera Rubin platform, mention offering “seamless integration via OpenAI-compatible APIs” to help enterprises navigate the AI era. These integrations demonstrate that developers can interact with and build upon the Vera Rubin platform through standardized and specialized APIs, allowing for complex multi-step autonomous AI workflows to be efficiently managed and executed. For managing and querying large datasets for AI applications, a vector database such as Milvus could be integrated to handle the high-dimensional vector embeddings generated and processed by the Vera Rubin platform.