Meta is seeking an experienced software engineer to join our Accelerator Solutions & Technologies group, supporting the development of Meta’s accelerators collective communications software library and optimizing distributed AI/ML workloads’ performance. This is an opportunity to work with a highly skilled engineering team, collaborating with a large set of cross-functional and international partners. Meta’s next-generation, super-cluster AI/ML platforms offer the opportunity to work in an extremely dynamic environment, enabling core technologies deployed in some of the world’s largest scale clusters.

MINIMUM QUALIFICATIONS:

• Bachelor’s degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
• Masters or PhD in Computer Science, Computer Engineering, or any other relevant technical field
• 6+ years experience in developing C++ codebase
• Understanding of performance, benchmarking measurements, and optimization of collective communication and distributed at-scale model training

PREFERRED QUALIFICATIONS:

• Understanding of the transport stack (e.g., RoCE), its constraints and performance measures and how transport considerations enable the collective communications stack.
• Knowledge of AI/HPC hardware requirements and specifications (e.g., configuring hardware components, GPU, memory, network for AI/HPC workloads).
• Full-stack experience and understanding of AI/HPC systems, from HW/infrastructure through the application layer, performance optimizations, including familiarity with relevant tools, libraries, and frameworks (e.g., NCCL, PyTorch, CUDA).
• Experience in one or more of the following machine learning domains: hardware accelerators, AI Infrastructure, and/or high performance computing (HPC), particularly pertaining to interconnect and collective.

• Understand and contribute to the collective communications library, intended to be deployed on Meta’s AI/ML superclusters
• Design and implement communication features for next generation AI/ML workloads
• Support networking and compute hardware acceleration techniques to improve ML inference and training model performance
• Support large-scale deployment of collective communication libraries across Meta’s infrastructure
• Perform architectural analysis to ensure system designs meet performance, scalability, and reliability requirements
• Analyze simulation results to guide firmware development and optimization efforts