Software Engineering Manager, AI Networking

Job Description

About the Role

In this role, you will be a member of the Network AI Software team and part of the bigger DC networking organization. The team develops and owns the software stack around collective communication libraries around Meta. At the high level, the team aims to enable Meta-wide ML products and innovations to leverage our large-scale training and inference fleet through an observable, reliable and high-performance distributed AI communication stack. Currently, one of the team's focus is on building customized features, SW benchmarks, performance tuners and SW stacks around PyTorch to improve the full-stack distributed ML reliability and performance (e.g. Large-Scale GenAI/LLM training) from the trainer down to the network communication layer. And we are seeking for leaders to work on the space of GenAI/LLM scaling reliability and performance.

Key Responsibilities

Lead and manage a networking related Software Engineering Team within the Network AI Software group.
Develop and optimize software stacks around collective communication libraries for distributed ML training and inference.
Work on building customized features, benchmarks, and performance tuning tools for PyTorch and related distributed ML systems.
Focus on improving the reliability and performance of large-scale GenAI/LLM training from the trainer to the network communication layer.

Requirements

BS or MS in Computer Science or related technical discipline or equivalent experience.
2+ years of experience managing a networking related Software Engineering Team.
Working knowledge of network transport stack such as RoCE (RDMA).
Experience with software development for Distributed and Embedded systems.
Experience recruiting and managing Software Engineers.
Experience with NCCL and distributed GPU reliability/performance improvement on RoCE/Infiniband.
Experience working with Deep Learning frameworks like PyTorch, Caffe2 or TensorFlow.
Knowledge of ML, deep learning and LLM.