Senior Solutions Architect, AI Performance Engineering

We are looking for a Solutions Architect with a performance engineering background who can help our most sophisticated Autonomous Vehicles and Robotics customers accelerate Physical AI workloads using NVIDIA's full-stack technologies! As part of the Automotive Solutions Architecture team, we work with some of the most innovative accelerated computing platforms focused on the development and test of Autonomous Vehicles. We dive deep into customer projects to solve performance bottlenecks. We use insights from workloads to guide next-generation NVIDIA hardware and software. If you are driven by innovation and ambition, this is the team for you!   What you'll be doing: • Engaging directly with key AV application engineers to understand the current and future problems they are solving. You will develop and improve fundamental parallel algorithms and data structures. You will provide efficient solutions using GPUs through library development and direct application contributions. • Collaborating closely with the architecture, research, libraries, tools, and system software teams at NVIDIA to influence the build of next-generation architectures, software platforms, and programming models. • Engaging in deep optimization of high-performance operators, involving but not limited to GPU kernel optimization, instruction-level tuning, and compiler optimization. These optimizations will directly support customers to use NVIDIA libraries such as cuDNN, cuBLAS, and CUTLASS and Open- source libs like DeepGEMM, FlashMLA, FlashAttention, Flashinfer, etc. • Improving communication for Physical AI-related distributed transformer workloads by employing communication tools developed by NVIDIA. These include NCCL, NCCL GIN, and NVSHMEM, as well as open-source solutions like DeepEP and NCCL EP. This demands in-depth study of interconnect topologies (NVLINK) and network protocols (InfiniBand/RoCE) to develop efficient data transfer strategies alongside techniques enabling compute-communication overlap. What we need to see: • BSc/MSc/PhD or equivalent experience in Computer Science, Electrical Engineering, Physics, Mathematics, or a related technical field. • 8+ years of hands-on validated ML/DL performance engineering experience with focus on improving GPU compute efficiency of training and inferencing workloads. • Experience with C, C++, or Python and proficiency with Linux. • Solid understanding of software development, programming techniques, and algorithms. • Strong mathematical fundamentals, including linear algebra and numerical methods. • Background in parallel programming and high-performance computing, with extensive knowledge of parallel architectures and methods for performance analysis and tuning. Experience in GPU programming is desirable. • Experience in distributed communication optimization is highly helpful. This involves familiarity with remote direct memory access, GPU interconnects, collective communication algorithms, and associated open-source libraries used in large-scale model training and inference. • Effective verbal/written communication, and technical presentation skills. Ability to communicate your ideas/code clearly through blog posts, GitHub, ppt. Ways to stand out from the crowd: • Prior experience in writing CUDA kernels, and experience with Nsight System and Nsight Compute. • Experience in comprehensive evaluation and improvement of full-stack systems within at least one of these areas: LLM and HPC. Having expertise ranging from operator-level through framework-level to algorithm-level optimization is strongly preferred. • Proven software engineering fundamentals and system architecture thinking, with the ability to build modules and lead engineering approaches in complex systems. Widely considered to be