Skip to main content
Graphcore

AI Platform Architect

Graphcore Austin, Texas, United States 1 day ago
Digital Design & SoC Architecture AI/ML Hardware

Architect AI rack-scale platforms across PCIe Gen 5/6 pathways, NVMe storage topologies, and RDMA fabrics.

Originally posted November 18, 2026 by the employer.

About the role

Graphcore is seeking an AI Platform Architect to design and oversee the infrastructure stack that powers distributed AI workloads across hardware, software, compute, network, and storage. The role focuses on architecting an AI rack-scale platform optimized for trillion-parameter LLM training and high-throughput inference spanning PCIe Gen 5/6 pathways, NVMe storage topologies, and RDMA fabrics.

What you'll do

  • Define the end-to-end architecture for highly clustered AI environments, ensuring data flow across parallel storage systems, AI compute nodes, and high-bandwidth network fabrics.
  • Direct AI workload scheduling and orchestration strategy using Kubernetes or Slurm to manage distributed training jobs, model check-pointing, and inference serving.
  • Profile and eliminate system-level bottlenecks across deep learning frameworks like PyTorch and DeepSpeed down to OS-level NUMA pinning and I/O scheduling.
  • Drive hardware-software co-design with software, firmware, and OS teams to leverage hardware capabilities, including ARM mesh interconnects (RNI, HNF, SNF) and merchant silicon features.
  • Define 3-to-5-year AI platform requirements covering processor, memory, storage, GPU, thermal, mechanical, BIOS, and manageability disciplines alongside TCO (OPEX/CAPEX) analysis.

What you'll need

  • 4+ years functioning as a Lead or Principal Architect for large-scale AI or machine learning platforms within systems engineering, cloud architecture, HPC, or hardware engineering.
  • Deep practical knowledge of distributed AI training and deployment frameworks, including data, tensor, and pipeline parallelism for LLM architectures.
  • Authoritative understanding of system interconnects, system-level bottlenecks, and data pathways, including PCIe Gen 5/6, NVMe namespaces, and RDMA (RoCEv2/InfiniBand) integration.
  • Experience with container orchestration platforms and infrastructure-as-code (IaC) tailored for GPU-heavy bare-metal and cloud environments.
  • Demonstrated ability to translate AI model requirements into platform engineering requirement specifications that guide and influence silicon designs.

Skills: PCIe Gen 5/6, NVMe, RDMA, clustering, AI platform

Market context
  • Employer posting date: 18 Nov 2026
  • First seen in our index: 18 Nov 2026 (our observation date)
  • Last confirmed on the employer's career site: 19 Nov 2026
  • Specialty: Hardware Systems Engineering (AI ML Hardware)
  • Comparable active jobs: 1210 at 121 employers, as of 19 Nov 2026
  • Related specialties (Software, Firmware & Systems family): 6861 active jobs at 180 employers

View all Graphcore jobs · Hardware design jobs · Austin jobs

How we counted: postings read from employer career sites; dates are our observations unless labelled as the employer's. How we count

Graphcore

Graphcore

Apply now
Austin, Texas, United States
On-site
1 day ago

Share this job