AI Enablement Engineer

Ampstek; Stockholms län; On-site; SEK 900,000 - 1,200,000; Full time

Job summary

Ampstek seeks an infrastructure-focused AI platform engineer to deploy VMware Private AI Foundation with NVIDIA on Kubernetes. You will configure NVIDIA AI Enterprise components, manage GPU sharing and expose GPU usage and LLM metrics via Prometheus/Grafana.

You will validate AI components on the VKS platform and run scalability tests with thorough documentation. The role requires hands-on experience with NVIDIA NIM, DCGM exporter, and Kubernetes, plus proficiency in LLM inference serving and

Qualifications

• NVIDIA AI Enterprise on Kubernetes, including the NVIDIA GPU Operator and NIM deployment.
• GPU sharing with MIG and/or vGPU time-slicing.
• GPU and workload monitoring with Prometheus and Grafana (including DCGM exporter).
• Hands-on LLM inference serving and performance testing (e.g., NIM, Triton or vLLM).

Responsibilities

• Deploy and configure VMware Private AI Foundation with NVIDIA and NVIDIA AI Enterprise components, including NVIDIA NIM microservices and NeMo Guardrails.
• Configure GPU sharing (MIG and/or vGPU time-slicing) so GPUs are used efficiently across workloads.
• Expose GPU utilisation and LLM workload metrics to Prometheus (for example via the DCGM exporter), with dashboards.
• Validate that the client's AI components run correctly on the Kubernetes (VKS) platform.
• Run scalability tests for GPU and storage clusters and document the results.

Skills

NVIDIA AI Enterprise on Kubernetes

GPU sharing MIG/vGPU

Prometheus/Grafana monitoring

LLM inference serving

Tools

NVIDIA NIM deployment

DCGM exporter

Prometheus

Kubernetes (VKS)

Job description

Employment Type : Employment Payroll Contract

Duration : 12+ Months

Key responsibilities:

• Deploy and configure VMware Private AI Foundation with NVIDIA and NVIDIA AI Enterprise components, including NVIDIA NIM microservices and NeMo Guardrails.
• Configure GPU sharing (MIG and/or vGPU time-slicing) so GPUs are used efficiently across workloads.
• Expose GPU utilisation and LLM workload metrics to Prometheus (for example via the DCGM exporter), with dashboards.
• Validate that the client's AI components run correctly on the Kubernetes (VKS) platform.
• Run scalability tests for GPU and storage clusters and document the results.

Must-have:

• NVIDIA AI Enterprise on Kubernetes, including the NVIDIA GPU Operator and NIM deployment.
• GPU sharing with MIG and/or vGPU time-slicing.
• GPU and workload monitoring with Prometheus and Grafana (including DCGM exporter).
• Hands-on LLM inference serving and performance testing (for example NIM, Triton or vLLM).

Nice to have:

• VMware Private AI Foundation with NVIDIA
• NeMo Guardrails
• VKS or vSphere with Tanzu
• NVIDIA AI infrastructure certification, CKA or CKAD

Not a fit:

• Data scientists or ML researchers focused on model development rather than infrastructure
• Kubernetes engineers without GPU experience