Meta's AI Training and Inference Infrastructure is growing exponentially to support ever increasing use cases of AI. This results in a dramatic scaling challenge that our engineers have to deal with on a daily basis. We need to build and evolve our network infrastructure that connects myriads of training accelerators like GPUs together. In addition, we need to ensure that the network is running smoothly and meets stringent performance and availability requirements of RDMA workloads. These workloads expect a loss-less fabric interconnect with minimal latency. To improve performance of these systems we constantly look for opportunities across stack: network fabric and host networking, communications lib and scheduling infrastructure.
Responsibilities
As an AI/HPC System Performance Engineer, you will lead multi-disciplinary teams to develop solutions for large scale training systems. You will assess trade-offs of various solutions and make pragmatic decisions. You will ensure timely milestone delivery with teamwork and close collaboration. You will be responsible for the overall performance of the communication system, including performance benchmarking, monitoring and troubleshooting production issues. You will define technical strategy and drive a multi-year roadmap to make progress towards the related objectives. You will work with cross-functional teams and provide guidance on the AI network architecture including topologies, transport, congestion control techniques.
Minimum Qualifications
To be successful in this role, you will need a Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience. You will also need experience with developing, evaluating and debugging host networking protocols such as RDMA. You will need 10+ years of experience in designing, deploying and operating networks. You will need experience with triaging performance issues in complex scale-out distributed applications. You will need an understanding of AI training workloads and demands they exert on networks.
Preferred Qualifications
Experience with developing communication libraries, such as Message Passing Interface, NCCL, and UCX is a plus. Understanding of RDMA congestion control mechanisms on InfiniBand and RoCE Networks is also a plus. Understanding of the latest artificial intelligence (AI) technologies is a plus. Experience with machine learning frameworks such as PyTorch and TensorFlow is a plus. Experience in developing systems software in languages like C++ is a plus.
Benefits
In addition to base compensation, Meta offers benefits. Learn more about benefits at Meta.
XML job scraping automation by YubHub
