In the course, IT professionals learn skills such as designing network architectures optimized for AI workloads; GPU optimization; building for high-performance generative AI network fabrics; and ensuring the security, sustainability and compliance of networks that support AI. They also learn to incorporate the unique business requirements of AI, such as trade-offs for cost optimization and power, and the matching of computing power and cloud needs to manage carbon use.
These skills are desperately needed by technology leaders and the enterprises they support. According to the Cisco AI Readiness Index, 90% of organizations are investing to try to overcome skills gaps in AI.
The technology list for the new CCDE-AI Infrastructure certification covers four high-level domains, according to Cisco:
- AI/machine learning, compliance, and governance: This section differentiates between various AI use cases, how they impact the network, and how that network should be designed. This domain covers designing for regulations regarding data sovereignty and data locality, energy use, and cost optimization.
- Network: This domain includes the properties and functions provided by the network, the differences between different connectivity models, and ensuring adequate bandwidth.
- Security: Security must be built into the infrastructure, not bolted on later. This domain includes secure networks and the process of securing AI applications with relevant web filtering techniques. It also includes AI-enabled tools for log analysis and correlation
- Hardware and environment: This domain requires a thorough knowledge of specific hardware used to run AI, enabling candidates to choose and use them appropriately. This domain also includes storage options and the use of various AI-enabled tools for log analysis and correlation.
The CCDE v3.1 AI Infrastructure certification is so new that it won’t be available for testing until Feb. 9, 2025, coinciding with Cisco Live Amsterdam. But interested network professionals can visit CCDE Unified Exam Topics for materials to help prepare for the exam.
IBM: Cloud Pak for Watson AIOps
For network admins running IBM systems, the company is offering IBM Certified Administrator – Cloud Pak for Watson AIOps. This certification demonstrates that a system admin has expertise in IBM Cloud Pak for Watson AIOps v3.2, including AI Manager, Event Manager, and Metric Manager. The certification covers planning, sizing, installation, daily management and operation, security, performance, configuration of enhancements (including fix packs and patches), customization and/or problem determination.
During exam development, the subject matter experts define all of the tasks, knowledge, and experience that an individual would need to successfully fulfill their role with the product. The exam is 90 minutes, 65 questions, and assumes a level of on-the-job experience.
