Meta is seeking a Machine learning SOC Engineer to streamline Architecture validation and ensure the optimal performance of their AI Accelerator by developing and maintaining the host runtime, device drivers, APIs, and libraries.
Requirements
- 8+ years of experience in development in the area of heterogeneous computing or high-performance computing
- Programming skills in languages such as C++, Rust, and Python
- Experience with runtime development, device driver development, API design, and library implementation
- Knowledge of computer architecture, operating systems, and parallel programming models (e.g., OpenMP, MPI)
- Familiarity with heterogeneous computing platforms, such as NVIDIA CUDA, OpenCL, or AMD ROCm
- Experience with object oriented, trait oriented and functional programming
- Knowledge of machine learning frameworks, such as TensorFlow or PyTorch
Responsibilities
- Design and develop the host runtime environment for heterogeneous computing systems, ensuring efficient communication and data transfer between different processing units
- Develop and implement device drivers, APIs, and libraries to enable seamless interaction between the host runtime environment and various processing units
- Collaborate with hardware engineers to understand the architecture of heterogeneous computing systems and optimize the host runtime environment accordingly
- Work closely with application developers to ensure that the host runtime environment meets their requirements and provides optimal performance for their applications
- Participate in code reviews, and stay up-to-date with industry trends and emerging technologies
- Troubleshoot and debug issues, working closely with hardware engineers and application developers to resolve problems
Other
- Problem-solving skills to analyze complex technical issues and develop effective solutions
- Communication and collaboration skills to work effectively with cross-functional teams
- Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
