Deep Learning Systems Software Engineer

As a senior engineer at NVIDIA, you will play a crucial role in the optimization of deep learning workloads on cutting-edge hardware and software platforms. Your primary responsibility will be to understand, analyze, and profile these workloads to achieve peak performance. By building automated tools for workload analysis and optimization, you will contribute to enhancing the efficiency of GPU utilization and cloud application performance across diverse GPU architectures. Collaboration with cross-functional teams will be essential as you identify bottlenecks and inefficiencies in application code, proposing optimizations to drive end-to-end platform optimization. Your role will involve designing and implementing performance benchmarks and testing methodologies to evaluate application performance accurately. To qualify for this role, you should hold a Master's degree in CS, EE, or CSEE, or possess equivalent experience. With at least 5 years of experience in application performance engineering, you are expected to have a background in deep learning model architectures, proficiency in tools such as NVIDIA NSight and Intel VTune, and a deep understanding of computer architecture and GPU fundamentals. Proficiency in Python and C/C++ will be essential for analyzing and optimizing application code effectively. To stand out from the crowd, strong fundamentals in algorithms and GPU programming experience (CUDA or OpenCL) will be highly beneficial. Hands-on experience in performance optimization and benchmarking on large-scale distributed systems, familiarity with NVIDIA's server and software ecosystem, and expertise in storage systems, Linux file systems, and RDMA networking will further distinguish you as a top candidate. Joining NVIDIA means being part of a dynamic team that leads the AI revolution, offering you the opportunity to directly impact the hardware and software roadmap in a fast-growing technology company. If you are unafraid to tackle challenges across the hardware/software stack and are passionate about achieving peak performance in deep learning workloads, we want to hear from you.,