英伟达Solutions Architect - CPU and LPU
任职要求
• MS or PhD in Computer Science, Engineering, Mathematics, Physics, or a related field, or equivalent experience, plus 5+ years in AI systems, infrastructure, performance engineering, or solution architecture. • Strong understanding of modern CPU architecture, Linux systems, and software performance tuning, along with hands-on experience in AI inference for LLM, generative AI, or agentic AI workloads. • Experience optimizing heterogeneous systems involving CPU and accelerators, with familiarity in frameworks such as PyTorch, Triton, TensorRT-LLM, vLLM, or ONNX Runtime. • Strong programming, problem-solving, and communication skills, wi…
工作职责
aNVIDIA’s Solutions Architect team is looking for a software-focused Solutions Architect to drive adoption of next-generation AI infrastructure across NVIDIA CPU platforms and LPU-based inference systems. This role will focus on NVIDIA CPUs, including Grace, Vera, and future CPU generations, and on LPU platforms and LPX-class systems used to accelerate large language model inference and other latency-sensitive generative AI workloads. We are looking for someone who understands that AI efficiency is a full-stack challenge spanning model architecture, runtime, compiler, serving framework, host software, memory movement, and workload partitioning across CPU, GPU, and LPU.As a Solutions Architect, you will be the first line of technical expertise between NVIDIA and our customers for CPU- and LPU-centric AI system design. You will help customers understand how NVIDIA CPUs and LPU-based systems can improve the efficiency, latency, throughput, and total cost of their AI workloads, especially when deployed alongside NVIDIA GPUs in heterogeneous production environments. Your work will range from proof-of-concept development and software stack optimization to technical leadership with customer architects, engineering teams, and senior decision makers. You will engage directly with developers, ML engineers, researchers, platform architects, and IT leaders to identify bottlenecks, design optimization strategies, and build deployable reference architectures. You will also work closely with NVIDIA engineering, product, and field teams to translate customer needs into platform feedback, solution patterns, and roadmap inputs. What you’ll be doing: • Evangelize NVIDIA CPU platforms, including Grace, Vera, and future generations, as well as LPU-based systems and LPX-class platforms, with a strong focus on AI software stacks and workload efficiency. • Help customers design and optimize AI workloads across CPU, GPU, and LPU, improving latency, throughput, utilization, and overall cost efficiency. • Analyze and tune LLM and generative AI pipelines across serving, runtime, memory, I/O, batching, scheduling, and orchestration layers. • Build proof-of-concepts, reference architectures, and technical guidance in partnership with Engineering, Product, and Sales teams. • Establish trusted technical relationships with customer architects, infrastructure teams, and senior leaders, becoming a strategic advisor for heterogeneous AI system design.
• Analyze the performance of a wide range of machine learning and deep learning algorithms across existing and emerging architectures. • Identify bottlenecks and devise creative software solutions or recommend improvements in GPU architectures. • Explore and evaluate how hardware and software architectures interact with future algorithms and applications.
• Work with Sales, BD and CPM team to introduce NVIDIA technologies into assigned accounts and grow business accordingly. • Serve as the primary technical authority on CPU technologies for NVIDIA’s Chinese CSP customers, providing expert consultation on CPU selection, architecture design, and integration with NVIDIA’s AI infrastructure (including Grace/Vera CPUs and NVL72 platforms). • Lead CPU-focused technical engagements with CSPs, collaborating with their R&D, infrastructure, and AI teams to understand workload requirements (e.g., AI data preprocessing, HPC, distributed computing) and design optimized CPU-GPU integrated solutions. • Drive CPU performance optimization for CSP workloads, conducting in-depth analysis of bottlenecks, implementing tuning strategies (including SIMD instruction set optimization and low-level intrinsics), and delivering reference implementations to unlock full platform potential. • Act as a liaison between CSP customers and NVIDIA’s global engineering, product, and R&D teams, advocating for customer-specific CPU requirements, providing feedback on product roadmaps, and ensuring alignment with NVIDIA’s technical strategy and export compliance guidelines. • Lead technical workshops, training sessions, and proof-of-concept (PoC) projects for CSPs, demonstrating the value of NVIDIA’s CPU-integrated solutions and enabling customer teams to effectively leverage these technologies. • Monitor industry trends in CPU technology, data center architectures, and CSP workload evolution, providing strategic insights to internal teams to enhance NVIDIA’s CPU-related products and solutions for the Chinese market. • Mentor junior technical team members, share CPU expertise, and drive best practices in CSP technical engagement and solution delivery.
• Lead presales and architecture engagements with AI industry customers, focusing on GPU servers, AI clusters, and large‑scale training/inference platforms built on NVIDIA HGX, GPU systems, and reference architectures. • Design and validate end‑to‑end AI data center solutions, including server platforms, storage connectivity, and high‑performance networking based on Spectrum, Quantum, ConnectX, and BlueField. • Define system architectures for AI supercomputing, LLM training, and inference workloads, including node configuration, GPU topology, PCIe/NVLink considerations, and network design. • Support business teams in exploring, developing, and deploying NVIDIA server and GPU solution opportunities, from early technical discovery through POC and production rollout. • Own and execute POCs and hands‑on labs that validate GPU server performance, scalability, reliability, and interoperability across compute, storage, and network domains. • Troubleshoot complex end‑to‑end issues involving GPU servers, firmware, drivers, operating systems, and networking stacks, and drive fixes with internal R&D and partners. • Provide structured feedback on platform features, system requirements, and customer needs to server OEMs, engineering, and product teams to improve NVIDIA AI platforms and ecosystems.
THE ORGANIZATION: The Physical AI and Vertical Software & Solutions (PAVS) organization is dedicated to accelerating the adoption of Physical AI across robotics, industrial, and automotive markets. PAVS delivers best-in-class vertical software and co-engineered solutions that shorten design cycles for lighthouse customers while improving solution quality and user experience. THE ROLE: This position offers an exciting opportunity to work on the cutting edge of Physical AI solutions. As a senior technical contributor, you will play a key role in driving end-to-end delivery of customer software solutions. Your responsibilities will include defining customer reference architectures, contributing to solution design, and coordinating implementation and optimization across multiple teams.