Chief Architect, Multitenant GPU Cloud Clusters

Nava; Singapore; On-site; SGD 240,000 - 360,000; Full time

Job summary

Nava is building Asia’s next-generation AI-native cloud platform in Singapore to empower AI developers with scalable, high-performance infrastructure. We are scaling a cross‑functional team to design and operate multi‑tenant GPU clusters powering large‑scale AI training and inference, with deep integration to the Nava Data Platform.

This Head of Cluster Engineering role combines hands‑on engineering with strategic leadership, shaping the core systems that deliver low‑latency, reliable AI compute

Qualifications

• 10+ years in systems, infrastructure, or cluster engineering
• 5+ years in a senior leadership role
• Deep Linux kernel internals and GPU driver expertise
• Extensive NVIDIA ecosystem tools: CUDA, NCCL, NVLink/NVSwitch, DOCA, GPUDirect, TensorRT
• Proven success designing and scaling high‑performance, multi‑tenant GPU clusters for AI workloads
• Strong orchestration experience: Kubernetes, Slurm, or Volcano
• Experience optimizing high‑speed interconnects: InfiniBand, RoCE v2, RDMA, eBPF/Cilium
• Experience with GitOps, infrastructure-as-code and hardware telemetry observability

Responsibilities

• Lead and scale an elite engineering team focused on control plane architecture and cluster orchestration
• Direct automated bare‑metal provisioning, BIOS/UEFI settings, BMC/IPMI integrations and driver deployment
• Drive deep hardware‑software co-design with NVIDIA stack to maximize GPU utilization
• Develop topology‑aware scheduling and fabric‑aware resource allocation
• Optimize InfiniBand/RDMA networks and modern networking stacks to reduce latency
• Integrate infrastructure with Nava Data Platform and optimize data pipelines for large-scale training
• Establish GitOps pipelines, robust SRE practices, and hardware telemetry for health and performance

Skills

Engineering leadership

Linux kernel internals

GPU driver stack

NVIDIA ecosystem

Kubernetes orchestration

Slurm/Volcano scheduling

Networking (InfiniBand)

GitOps / IaC

Tools

CUDA/NCCL

NVLink/NVSwitch

TensorRT

DOCA / DPU

Job description

Nava is building Asia’s next-generation AI-native cloud platform in Singapore to empower AI developers with scalable, high-performance infrastructure. We are scaling a cross‑functional team to design and operate multi‑tenant GPU clusters powering large‑scale AI training and inference, with deep integration to the Nava Data Platform.

This Head of Cluster Engineering role combines hands‑on engineering with strategic leadership, shaping the core systems that deliver low‑latency, reliable AI compute