Responsibilities:

• Design end-to-end AI/ML platform architectures spanning inference, training, and data pipelines for enterprise customers
• Develop reference architectures for GPU cluster deployment, LLM serving, and multi-tenant ML infrastructure
• Advise on GPU fabric topology, including NVLink, InfiniBand, and RoCEv2, for distributed training environments
• Act as the primary technical advisor and escalation point for assigned customers, leading root cause analysis on complex production issues
• Deliver technical workshops, proof-of-concept engagements, and executive-level presentations on AI infrastructure strategy
• Design observability strategies across DCGM, OpenTelemetry, eBPF, and GPU metrics pipelines
• Partner with customer platform, MLOps, and data science stakeholders to translate workload requirements into scalable architecture
• Feed customer insights back into the product and engineering roadmap, and mentor junior members of the Solutions Architecture team

Skills/Must have:

• Experience: 8+ years in infrastructure, platform, or solutions engineering, including 3+ years focused specifically on AI/ML infrastructure or MLOps
• Core Tech/Domain: Deep, hands-on Kubernetes expertise (cluster lifecycle, workloads, operators, RBAC) plus direct experience with NVIDIA GPU infrastructure (H100/H200/B200 preferred)
• Methodology/Protocols: Working knowledge of distributed training concepts (NCCL, tensor and pipeline parallelism), LLM inference serving and optimisation (vLLM, NIM, TGI), and GPU networking (GPU Operator, MIG, SR-IOV, IB/RoCEv2)
• Soft Skills: Confident, credible communicator able to run technical discussions with engineers through to executive stakeholders; strong troubleshooting mindset and customer advisory presence