Senior Solutions Architect, Inference Service Providers

We are looking for an architect that wants to change how the AI inference industry works. With agent adoption taking off and GPUs in shortage, NVIDIA has a key part to play in solving the technical challenges that scale token generation on finite resources. Our ISP (inference service provider) team supports some of our most strategic partners who rely on SA support to push the boundaries of performance, squeezing value and reliability out of vast NCP ( NVIDIA Cloud Partner) infrastructure. You’ll master full-stack inference with every SOTA bell and whistle, from smarter routing to rack-scale disaggregation down to kernel optimization. You’ll write reference architectures adopted by 10k+ gpu clusters. You’ll teach C-level executives, distinguished engineers and datacenter developers how to scale their businesses, systems and infrastructure. This role covers a wide range of technical skills and the ability to lead high-stakes conversations with partners. Nobody can cover everything! For that reason, we're looking for someone with fast-learning technical agility and comfort with partners or customers. We’ll build a leading tech stack internally, then bake all our best tricks into a reference architecture and guide the industry to follow in our footsteps. What you'll be doing: Work with inference partners who build cutting-edge inference services, teach them the value of our stack and bring back product insights that benefit the whole industry. Build and operate inference recipes with tools like NVIDIA Dynamo, distributing tasks among GPU workers to improve efficiency. Accelerate inference pipelines using TensorRT-LLM, vLLM, SGLang, and other backends to ensure seamless integration with disaggregated inference. Evangelize devops best-practices for such as managing kubernetes clusters, configuring compute fabrics, and observability. Provide mentorship and technical leadership to customers and internal teams, guiding them through the deployment of disaggregated inference