NVIDIA seeks a Senior Networking (ETH/IB) Solutions Architect to join its NVIDIA Infrastructure Specialist Team. This role is instrumental in supporting academic and commercial groups revolutionizing deep learning, data analytics, and powering data centers with NVIDIA products. You will contribute to building some of the world's largest and fastest AI/HPC systems, working within a dynamic, customer-focused team that demands excellent interpersonal skills. Your responsibilities will involve deep engagement with customers, partners, and internal teams to analyze, define, and implement large-scale networking projects encompassing networking, system design, and automation, acting as a key customer interface.

Key duties involve building AI/HPC infrastructure for new and existing clients. You will support the operational and reliability aspects of large-scale AI clusters, emphasizing performance at scale, real-time monitoring, logging, and alerting. This role requires engagement throughout the entire service lifecycle, from inception and design to deployment, operation, and ongoing refinement. Maintaining live services includes measuring and monitoring availability, latency, and overall system health, while also providing crucial feedback to internal teams through bug reporting, workaround documentation, and improvement suggestions.

A Bachelor's or Master's degree (or equivalent experience) in Computer Science, Electrical/Computer Engineering, Physics, Mathematics, or a related field is required. Candidates need at least 5 years of professional experience in networking fundamentals, specifically within the Ethernet or InfiniBand domains. Essential hands-on experience includes working with network switch/router platforms such as Cumulus Linux, SONiC, IOS, JunosOS, and EOS. A solid understanding of Ethernet, InfiniBand, and RDMA core principles is vital. Proficiency in end-to-end IB/Eth cluster deployment, adapter configuration, and firmware maintenance is expected, along with the ability to conduct professional performance benchmarking using mainstream RDMA testing tools.