Lead AI Platform Architect (GPU Compute & Secure AI Services)

Title: Lead AI Platform Architect (GPU Compute & Secure AI Services)

Location: Hoofddorp (On-site, 5 days/week)

Type: Full-time

Language: Fluent English required (Dutch not required)

Eligibility: Must live and work in the Netherlands

About the role

We sell AI hardware and operate our own infrastructure. We are building a new GPU compute platform and a set of secure AI service capabilities on top of it for enterprise use cases.

This is a greenfield initiative: you will own the end-to-end architecture across physical design, platform software layers, and the roadmap required to reach product-market fit. You will lead this initiative independently, reporting to the Head of Technology and working closely with other stakeholders.

What you’ll own

Platform and infrastructure

• Define the reference architecture for GPU-based compute (compute, networking, storage, security boundaries, capacity planning, operational standards).
• Design a production-grade, multi-tenant platform layer built around Kubernetes, including GPU scheduling/isolation, identity and access management, tenancy boundaries, policy controls, and automation.

Searching for IT & Network Engineering roles that provide visa sponsorship? Connect with international employers through IT & Network Engineering Jobs with Visa Sponsorship opportunities actively seeking talented professionals.

• Establish observability and reliability practices: metrics, logs, traces, SLOs/SLAs, incident readiness, upgrades, and lifecycle management.
• Deliver the first production deployment in a single datacenter, with a clear path to scale out to multiple datacenters soon after.

Secure AI service capabilities (confidential initiative)

• Architect and guide implementation of a service layer for model serving and AI workloads on our infrastructure, including:
• High-performance inference serving patterns (for example vLLM-class approaches)
• Retrieval-Augmented Generation pipelines (chunking, embeddings, retrieval quality, governance and auditability)
• Vector search and knowledge storage (for example Qdrant-class systems)
• Strict tenant isolation and data residency controls from day one
• Drive buy vs build decisions for platform components and define the integration architecture, with freedom to propose alternatives.

Team and execution

• Start architecture-led, then progressively become more hands-on as the platform moves into build and production phases.
• Define the hiring plan for the initiative: which roles we need, sequencing, and what to outsource versus build in-house.
• Mostly internal role, with the ability to join customer conversations when deep technical credibility is needed.
• Support an initial assisted onboarding approach for v1, while designing the platform so we can move to self-service provisioning quickly.

What we’re looking for

• Senior experience designing and operating cloud platforms or infrastructure products.
• Strong production experience with GPU workloads and container platforms.
• Ability to lead ambiguous greenfield initiatives, make pragmatic trade-offs, and drive execution with stakeholders.
• Comfortable defining a multi-tenant architecture from day one (hard isolation per customer).
• Experience with NVIDIA GPU infrastructure/software stack and operating GPU clusters at scale is a plus.
• Experience supporting enterprise customers and multi-tenant platforms is a plus (not a strict requirement).

Interested in opportunities specifically in Netherlands? Discover our dedicated Visa Sponsorship Jobs in Netherlands page featuring roles from top employers in this location.

• Fluent English proficiency; Dutch is not required.
• Must be able to live and work in the Netherlands (relocation and visa sponsorship available).

Preferred experience (nice to have)

• Experience designing inference-serving platforms (request routing, batching, cost/utilization optimization).
• Building RAG systems end-to-end: embedding models, chunking strategy, retrieval evaluation, governance/auditability.
• Exposure to metering/chargeback and integrating with existing billing solutions.
• Multi-datacenter platform design considerations (capacity, rollout strategy, operational model).
• Experience designing self-service provisioning (portal/API) for infrastructure platforms.

Working model

This is an on-site role in Hoofddorp, 5 days per week, to stay close to the hardware, platform engineering, and execution environment.