IBM Research is seeking a Research Scientist to work at its T. J. Watson Research Center in Yorktown Heights, NY.
You will be expected to be proficient in C, C++ and Python programming, and must be well versed in state-of-the-art software collaboration tools, to design and implement the software stack for new AI hardware accelerators. You must demonstrate a passion for best practices in software engineering with the ability to work within a diverse team including researchers in exploratory accelerator architectures, compilers, and AI applications. Experience with the use and adaptation of AI frameworks including TensorFlow and PyTorch as well as CUDA for GPU programming is highly desirable. As a member of the software team, your responsibilities will include managing the end-to-end CI/CD of the software stack and conducting performance measurements of the Hardware/Software stack for standardized AI benchmarks.
As hardware/software co-design for specialized hardware accelerators is one of the hallmarks of the project, in addition to proven software engineering skills, a background in architecture/microarchitecture and knowledge of AI models is preferred. As a member of the Research team, you will need to engage in research topics and cover new domains quickly, apply high standards to developing and nurturing impactful research projects in complex unexplored domains.
* At least 2-5 years experience in C++ and Python programming
* Experience with software development best practices, collaboration tools, CI/CD and DevOps
* Experience in use and adaptation of deep learning frameworks such as PyTorch, TensorFlow, TVM and/or ONNX runtimes
* Master's degree in Computer Science or a related area
* Knowledge of deep learning and machine learning concepts, and an aptitude to work with rapidly evolving state-of-the-art models
* Knowledge of compilers, computer architecture, and specialized hardware accelerators
* Knowledge of cloud infrastructure and containerized deployment