高级解决方案工程师 – GPU 与 AI 基础设施
Senior Solution Engineer – GPU & AI Infrastructure
#### **高级解决方案工程师 – GPU 与 AI 基础设施**
* * *
**关于 Civo：**
Civo 是一家高性能的新云服务提供商，专为现代 AI、高性能计算（HPC）和云原生基础设施的需求而设计。我们消除传统云的冗余开销，提供超低延迟计算、裸金属 GPU 性能以及可扩展的 Kubernetes 编排。
专为 AI 工程团队、企业及研究机构设计，Civo 提供直接访问前沿 NVIDIA GPU 集群、高速网络架构和并行存储系统的权限，以高效地训练、微调和部署基础模型。我们将高密度基础设施与可预测的定价和最大计算吞吐量相结合，使组织能够在不增加复杂性或成本负担的情况下扩展 AI 工作负载。
* * *
**职位描述：**
作为高级解决方案工程师 – GPU 与 AI 基础设施，您将担任 Civo 大规模 AI 和高性能计算（HPC）客户项目的首要技术架构师。您将负责设计用于训练和推理大规模基础模型的先进 NVIDIA GPU 集群。
在该职位中，您将弥合客户业务目标与超高性能硬件执行之间的差距。您将主导技术对接，将复杂的 AI 工作负载需求转化为生产就绪的高层设计（HLD）、低层设计（LLD）和详细物料清单（BOM）。您的专业知识将涵盖基于裸金属和 Kubernetes 的编排，适用于前沿 NVIDIA Blackwell 架构（如 B300 和 GB300NVL），使用超低延迟 InfiniBand 和高速 RoCE 网络架构。
* * *
**职责：**
**解决方案设计与架构**
- 系统设计文档：撰写企业级 GPU 超算集群的全面高层设计（HLD）和低层设计（LLD）文档。
- 物料清单（BOM）：生成涵盖计算节点、NVLink 交换机、网络架构、收发器/线缆、液冷/风冷需求、电力分配和高性能存储的详细 BOM。
- GPU 集群拓扑：为 NVIDIA Blackwell 平台（特别是 B300 和 GB300NVL 机架级架构）设计扩展型（NVLink/NVSwitch）和扩展外（Fat-Tree、Rail-Optimized）网络拓扑。
- 网络架构工程：设计高吞吐、低延迟的网络架构。
查看英文原文
#### **Senior Solution Engineer – GPU & AI Infrastructure**
* * *
**About Civo:**
Civo is a high-performance neocloud provider purpose-built for the demands of modern AI, high-performance computing (HPC), and cloud-native infrastructure. We eliminate legacy cloud overhead to deliver ultra-low-latency compute, bare-metal GPU performance, and streamlined Kubernetes orchestration at scale.
Purpose-designed for AI engineering teams, enterprises, and research institutions, Civo delivers direct access to cutting-edge NVIDIA GPU clusters, high-speed fabrics, and parallel storage systems required to train, fine-tune, and deploy foundation models efficiently. We combine high-density infrastructure with predictable pricing and maximum compute throughput, empowering organizations to scale AI workloads without the complexity or cost bloat of traditional hyperscalers.
* * *
**About the Role:**
As a Senior Solution Engineer – GPU & AI Infrastructure, you will serve as the primary technical architect for Civo’s large-scale AI and high-performance computing (HPC) customer initiatives. You will be responsible for designing state-of-the-art NVIDIA GPU clusters tailored for training and inferencing massive foundation models.
In this role, you will bridge the gap between customer business objectives and ultra-high-performance hardware execution. You will lead technical engagements, translate complex AI workload requirements into production-ready High-Level Designs (HLD), Low-Level Designs (LLD), and detailed Bills of Materials (BOM). Your expertise will span bare-metal and Kubernetes-based orchestrations across cutting-edge NVIDIA Blackwell architectures (e.g., B300 and GB300NVL) using ultra-low-latency InfiniBand and high-speed RoCE networking fabrics.
* * *
**Responsibilities:**
**Solution Design & Architecture**
- System Design Documents: Author comprehensive High-Level Design (HLD) and Low-Level Design (LLD) documentation for enterprise-scale GPU supercomputing clusters.
- Bill of Materials (BOM): Generate detailed BOMs covering compute nodes, NVLink switches, network fabrics, transceivers/cabling, liquid/air cooling requirements, power distribution, and high-performance storage.
- GPU Cluster Topology: Architect scale-up (NVLink/NVSwitch) and scale-out network topologies (Fat-Tree, Rail-Optimized) for NVIDIA Blackwell platforms, specifically B300 and GB300NVL rack-scale architectures.
- Fabric & Networking Engineering: Design high-throughput, low-latency networking architectures utilizing both InfiniBand (e.g., NDR/X800) and RoCE / RoCEv2 (e.g., NVIDIA Spectrum-X / Spectrum-4) with lossless Ethernet mechanisms (PFC, ECN, Adaptive Routing).
- Multi-Tenant & Deployment Models: Deliver tailored architectures for both Bare-Metal (Slurm, OpenMPI, bare-metal provisioning) and Cloud-Native / Kubernetes environments (NVIDIA GPU Operator, Network Operator, Run:ai, KubeFlow).
- Storage Integration: Architect high-bandwidth parallel storage solutions utilizing GPUDirect Storage (GDS) and enterprise AI file systems (e.g., VAST Data).
**Technical Sales Support & Customer Engagement**
- Partner with Civo’s sales and commercial teams as the technical lead for high-value AI infrastructure opportunities.
- Engage directly with customer CTOs, Chief AI Officers, infrastructure leads, and ML engineers to evaluate technical requirements, compute sizing, and fabric choices.
- Lead deep-dive architectural workshops and technical presentations on Civo's bare-metal GPU and managed Kubernetes offerings.
- Produce precise technical proposals and lead responses to complex RFPs/RFIs regarding AI infrastructure.
**Proof-of-Concept (PoC) & Benchmarking**
- Architect and oversee Proof-of-Concept (PoC) deployments to validate real-world performance for customer workloads.
- Benchmark cluster performance using industry-standard tools (NCCL tests, GPUDirect RDMA latency/bandwidth, MLPerf, Megatron-LM benchmarks).
- Address network congestion, fabric routing, and thermal/power optimization during validation phases.
**Product & Ecosystem Collaboration**
- Serve as the bridge between enterprise AI clients, hardware vendors (NVIDIA, network OEMs), and Civo’s internal platform engineering team.
- Provide continuous feedback to product teams on market trends, hardware platform demands, and feature requirements for AI/GPU orchestration.
* * *
**Key Results/Objectives:**
- Technical Wins: Achieve high technical win rates on large-scale AI/GPU cluster sales opportunities.
- Design Excellence: Successfully deliver complete, peer-reviewed HLDs, LLDs, and BOMs within target deal timelines.
- Customer Satisfaction: Achieve successful PoC completion and sign-off for enterprise clients scaling AI workloads on Civo infrastructure.
* * *
**Requirements:**
#### **Experience & Core Qualifications**
- 5+ years in a Solution Architecture, Systems Engineering, or Technical Pre-Sales role focused on high-performance cloud, HPC, or AI infrastructure.
- Bachelor’s degree in Computer Science, Electrical Engineering, Systems Engineering, or equivalent practical experience.
#### **Technical Expertise**
- NVIDIA GPU Architecture: Deep hands-on knowledge of NVIDIA HGX/DGX platforms, NVLink/NVSwitch fabrics, and Blackwell architectures (B300, GB300NVL, GB200 NVL72/NVL36).
- High-Speed Networking: Expert-level knowledge of cluster fabric topologies:
- InfiniBand: Quantum-2 / Quantum-X800, Subnet Management, Adaptive Routing.
- RoCE / RoCEv2: Spectrum-X / Spectrum-4 Ethernet switches, PFC, ECN, RoCE configuration, and optimization.
- GPU Direct Technologies: GPUDirect RDMA (GDR) and GPUDirect Storage (GDS).
- Orchestration & Platforms: Proficiency in deploying and optimizing GPU workloads on:
- Kubernetes: Container networking (CNI), NVIDIA GPU Operator, RDMA Shared Device Plugin, MPI Operator.
- Bare-Metal: Slurm, Ansible, Terraform, PyTorch/NCCL environment tuning.
- Documentation Skills: Demonstrated experience creating enterprise-grade HLDs, LLDs, network rack diagrams, and itemized BOMs.
- Power & Thermal Awareness: Familiarity with high-density datacenter environments, liquid cooling technologies (Direct-to-Chip, CDU/liquid loop setups), and power delivery constraints for 100kW+ per rack deployments.
#### **Soft Skills**
- Strong technical leadership and presentation skills, with the ability to articulate complex network and hardware tradeoffs to executive stakeholders.
- Problem-solving mindset capable of diagnosing complex hardware-software interaction bottlenecks in distributed training/inference setups.
#### **Location**
- Must be UK based.
* * *
**Nice to Have:**
- NVIDIA Certified Professional: AI Infrastructure (NCP-AII).
- NVIDIA Certified Professional: AI Networking (NCP-AIN).
- NVIDIA Certified Professional: InfiniBand (NCP-IB).
- NVIDIA Certified Associate / Professional: AI Workload Deployment & Cloud Native.
* * *
**Why Join Civo?**
- Competitive compensation and benefits package.
- 4-day week company (unless attending an event).
- Uncapped holiday.
- Remote work environment with flexibility and autonomy.
- Collaborative and inclusive culture that values diversity and creativity.
- Opportunity to work with a dynamic and innovative team in the fast-growing cloud industry.