Senior System Software Architect, HPC and AI Networking
Core
Design and prototype scalable software systems optimizing distributed AI training and inference, focusing on throughput, latency, and memory efficiency across large-scale systems.
Role type
Senior System Software Architect (HPC and AI Networking)
Builds
Communication libraries (NCCL, UCX, UCC), runtime systems, and AI-specific protocol layers for deep learning workloads.
Domain
High-Performance Computing (HPC) and Artificial Intelligence (AI) Infrastructure
Deliverable
production ML models
Required skills
Distributed AI training and inference optimization, Communication library development (NCCL, UCX, UCC), AI framework integration (TensorFlow, PyTorch, JAX), Hardware co-design (GPUs, DPUs, interconnects), Algorithm design, System programming, Computer architecture, RDMA/RoCE networking, CUDA programming
Preferred skills
Designing communication middleware for HPC, Background with emerging GPU architectures, Influencing within dynamic matrix environments
Technologies
NCCL, UCX, UCC, TensorFlow, PyTorch, JAX, RDMA, RoCE, CUDA, GPUs, DPUs
Responsibilities
Design scalable software for distributed AI workloads, Develop and evaluate communication library enhancements, Collaborate with AI framework teams, Co-design hardware features for data movement, Contribute to runtime system evolution, Collaborate with customers for innovative solutions
Seniority
Senior, hands-on IC
