Senior HPC and AI Network Software Architect
Core
Architect scalable software systems for distributed AI training and inference, focusing on throughput, latency, resiliency, and memory efficiency across cluster-scale deployments.
Role type
Senior IC HPC and AI Network Software Architect
Builds
Scalable AI infrastructure, communication libraries (NCCL, UCX, UCC), and runtime systems for frontier AI workloads
Domain
High-performance computing, AI infrastructure, distributed systems
Deliverable
production ML models
Required skills
Systems programming, parallel/distributed computing, high-performance networking, C++, Python, CUDA, AI frameworks (PyTorch, TensorFlow, JAX), hardware-software co-design
Preferred skills
NCCL/UCX/UCC expertise, RDMA, collective communications, congestion-aware transport, LLM deployment patterns (sharding, pipelining)
Technologies
NCCL, UCX, UCC, CUDA, PyTorch, TensorFlow, JAX, GPUs, DPUs
Responsibilities
Build and evolve architecture for distributed AI training/inference; Develop next-gen communication/runtime capabilities in libraries; Partner with AI framework teams to build integrations; Collaborate on hardware/system-level features for GPUs/DPUs/interconnects; Drive innovation in runtime systems and protocol layers
Seniority
Senior, hands-on IC
