Nvidia Senior CUDA Mathematical Libraries Engineer 

United States, California 

113316817

In this role you will be interacting and collaborating with internal partners, industry professionals, and members of the scientific community to define and implement highly optimized new numerical algorithms. Your responsibilities include the design of public APIs, implementation of new algorithms, performance tuning and analysis, and other general software engineering work. You will need great interpersonal and communication skills and a strong focus on delivering high-quality and performance software.

What you’ll be doing:

• Prototype, develop, and optimize software for deep learning, data analytics, scientific research, and engineering applications that use high-performance computing.

• Analyze the performance of GPU or CPU implementations.

• Collaborate with internal and external partners to understand software use cases and requirements.

• Capitalize on opportunities for adding new functionality or for improving performance.

• Attend conferences and engage with the larger community on relevant topics

What we need to see:

• PhD in Computer Science, Applied Math, or related science or engineering field of study with 3+ years of experience. or a Masters or Bachelors degree in those fields with additional experience.

• Excellent C/C++ programming and software design skills, including debugging, performance analysis, and test design.

• Experience developing parallel numerical software libraries on HPC platforms using multi-threading and/or MPI.

• Deep understanding of dense linear algebra and computations in science, engineering, or deep learning.

• Ability to work independently, define project goals and scope, and manage your own development effort.

• Good teamwork, communication, and documentation habits.

Ways to stand out from the crowd:

• Strong fundamentals in algorithms.

• Prior experience with low level programming in assembly and performance tuning.

• GPU programming experience (CUDA or OpenCL) is a strong plus but not required.

• Good knowledge of CPU and/or GPU hardware architecture.

• A scripting language, preferably Python.