Embedded & Systems Software
Deploy AMD Instinct GPU clusters from delivery to sustained production excellence.
Originally posted August 27, 2026 by the employer.
About the role
As a Cloud and Customer Solutions Engineer on AMD's Applied AI team, you will embed directly with AMD's most strategic AI customers to take AMD Instinct GPU clusters from delivery to sustained production excellence. This role involves writing production code, operating live clusters, and carrying accountability for customer production outcomes.
What you'll do
- Own customer deployments end-to-end: cluster bring-up and burn-in, production readiness certification, workload onboarding, performance validation, and sustained production operation on AMD Instinct GPU fleets.
- Deploy and tune large-scale training and inference stacks (ROCm, vLLM, SGLang, RCCL, Kubernetes, Slurm) against customer-specific workloads and SLOs across cloud, NeoCloud, and bare-metal environments.
- Lead root-cause analysis and resolution of production incidents on customer clusters, including Sev-1 response, and drive fixes to permanent closure.
- Build observability, benchmarking, and validation tooling needed to certify clusters as production-ready.
- Transfer operational capability to customer teams through documentation, runbooks, and hands-on enablement.
- Convert field findings into upstream contributions (ROCm issues and patches, serving-framework improvements, reference-architecture updates) and provide structured field signal to AMD product, software, and silicon teams.
What you'll need
- 5+ years of production software or infrastructure engineering, including significant time operating or deploying systems in environments you did not build.
- Hands-on experience with GPU compute at scale: cluster deployment, distributed training or high-throughput inference, performance debugging, and workload optimization.
- Strong working knowledge of the modern AI infrastructure stack: Kubernetes and/or Slurm, containerized GPU workloads, collective communication libraries (RCCL/NCCL), high-performance networking (RoCE/InfiniBand), and observability tooling (Prometheus, Grafana).
- Cloud platform depth (AWS, Azure, GCP, or NeoCloud environments), including hybrid and bare-metal deployment patterns.
- Proficiency in Python and at least one systems language; comfort navigating and modifying large codebases you did not write.
- Working familiarity with LLM application patterns (inference serving, RAG, and agentic workflows) sufficient to deploy and troubleshoot them in customer environments.
- Direct customer-facing experience: embedded deployments, technical escalations, on-site engagements, or equivalent.
Nice to have
- Experience with ROCm and AMD Instinct GPUs strongly preferred; deep CUDA-ecosystem experience with demonstrated ability to work cross-platform also valued.
- Open-source contribution history in AI/ML infrastructure projects is a plus.
- Bachelor's or Master's degree in Computer Science, Computer Engineering, Electrical Engineering, or equivalent practical experience.
Skills: GPU clusters, ROCm, Kubernetes, Slurm, distributed training, inference
Market context
- Employer posting date: 27 Aug 2026
- First seen in our index: 28 Aug 2026 (our observation date)
- Last confirmed on the employer's career site: 21 Nov 2026
- Specialty: Infrastructure/Platform Software (GPU Software Stack)
- Comparable active jobs: 341 at 26 employers, as of 21 Nov 2026
- Related specialties (Software, Firmware & Systems family): 6876 active jobs at 180 employers
How we counted: postings read from employer career sites; dates are our observations unless labelled as the employer's. How we count