APAC Forward-Deployed GPU Infra Engineer

Hyperbolic; United States; Remote; USD 120,000 - 180,000; Full time

Job summary

Hyperbolic Labs is seeking a Technical Support Engineer to embed with reserved customers running multinode GPU clusters. You will stand up their clusters, hand them over, and stay involved to diagnose deep issues across hardware, fabric, and software.

You'll own environments, coordinate with providers to prove fault locations, and build runbooks for L1 teams. Strong Linux, NCCL, and GPU experience are essential for this customer-facing role.

Qualifications

• Cluster stand-up and handoff with documentation for the customer's engineers.
• Deep escalations: multinode and NCCL performance, GPU faults, drivers, fabric issues.
• Provider escalation with evidence to drive resolution and clear customer updates.
• Embedded ownership of named accounts: learn workloads, guide changes.
• Proactive monitoring and alerting using Grafana/Prometheus.
• Automate repeat tasks and create runbooks for L1 teams.
• Deep Linux experience and CLI fluency in complex environments.
• Hands-on multinode GPU experience: NCCL, InfiniBand or RoCE, topology/placement.
• Hardware fault triage on GPU nodes: XID, ECC, dmesg, nvidia-smi.
• Experience provisioning and operating GPU clusters with Kubernetes/Slurm.
• Clear, confident communication with customers, including bad news when needed.
• Judgment on when to dig deeper or pivot.

Responsibilities

• Stand up, validate, and hand over new customer clusters with proper docs.
• Handle deep escalations for multinode/NCCL performance and hardware faults.
• Coordinate with providers, bringing evidence to resolve faults.
• Own named customer accounts and learn their workloads to guide changes.
• Monitor clusters proactively using Grafana/Prometheus to catch faults early.
• Automate repetitive tasks and build runbooks for L1 teams.
• Maintain deep Linux expertise and CLI fluency in diverse environments.
• Apply hands-on multinode GPU experience with NCCL, InfiniBand/RoCE and topology.
• Triage hardware faults on GPUs: XID, ECC, dmesg, lspci, nvidia-smi.
• Provision GPU clusters with Kubernetes/Slurm and CI/CD tooling.
• Communicate honestly with customers, delivering bad news when necessary.

Skills

Cluster stand-up

Escalation management

Customer-facing

Proactive monitoring

Linux CLI

GPU expertise

Tools

NCCL

InfiniBand

RoCE

Grafana

Prometheus

Kubernetes

Slurm

Terraform

Ansible

nvidia-smi

Job description

Hyperbolic Labs is seeking a Technical Support Engineer to embed with reserved customers running multinode GPU clusters. You will stand up their clusters, hand them over, and stay involved to diagnose deep issues across hardware, fabric, and software.

You'll own environments, coordinate with providers to prove fault locations, and build runbooks for L1 teams. Strong Linux, NCCL, and GPU experience are essential for this customer-facing role.