AI/HPC Systems Performance Engineer

Meta•Menlo Park, CA

About The Position

Meta's AI Training and Inference Infrastructure is growing exponentially to support ever increasing uses 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 that expects a loss-less fabric interconnect. To improve performance of these systems we constantly look for opportunities across stack: network fabric and host networking, comms lib and scheduling infrastructure.

Requirements

Currently has, or is in the process of obtaining a Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience. Degree must be completed prior to joining Meta
Bachelor's degree in Computer Science, Computer Engineering, or other relevant technical field, with 2+ years work experience
Experience with using communication libraries, such as MPI, NCCL, and UCX
Experience with developing, evaluating and debugging host networking protocols such as RDMA
Experience with triaging performance issues in complex scale-out distributed applications

Nice To Haves

Understanding of AI training workloads and demands they exert on networks
Understanding of RDMA congestion control mechanisms on IB and RoCE Networks
Understanding of the latest artificial intelligence (AI) technologies
Experience with machine learning frameworks such as PyTorch and TensorFlow
Experience in developing systems software in languages like C++

Responsibilities

Collaborate with hardware and software teams to optimize end-to-end communication pathways for large-scale distributed training workloads, ensuring seamless integration between compute, storage, and networking components
Design, implement, and validate new collective communication algorithms tailored for AI/HPC workloads, leveraging RDMA and advanced networking technologies to maximize throughput and minimize latency
Develop and maintain automated performance testing frameworks for continuous benchmarking of communication libraries and RDMA transport layers, enabling rapid identification of regressions and bottlenecks
Analyze and profile communication patterns in real-world training jobs, using telemetry and tracing tools to uncover inefficiencies and recommend architectural improvements
Drive adoption of best practices for scalable, fault-tolerant communication in production environments, including tuning RDMA parameters, optimizing network fabric configurations, and ensuring robust error handling
Work closely with vendors and internal teams to evaluate and integrate new hardware features (e.g., NICs, switches, accelerators) that can enhance communication performance for AI/HPC clusters
Contribute to documentation and knowledge sharing by authoring technical guides, performance reports, and internal wiki pages to educate peers and stakeholders on communication system optimizations
Participate in code reviews and design discussions to ensure high-quality, maintainable solutions that meet the evolving needs of large-scale AI/HPC infrastructure