Skip to content

Capability: AI & Data

Intelligence you own, inside your own perimeter.

We design and build the AI and data platforms Saudi institutions can run themselves — GPU infrastructure, governed data foundations, and use cases that hold up to a regulator's questions.

  • Sovereign deployment
  • GPU infrastructure
  • Governed data
  • Auditable AI

What we build

Four workstreams, scoped independently and sequenced to your mandate.

  • AI infrastructure & GPU platforms

    Sizing, reference architecture and deployment of GPU compute — bare metal or virtualised — with the networking, storage and power assumptions made explicit before anything is procured.

  • Data platforms & governance

    Lakehouse and warehouse design, data classification aligned to PDPL, lineage, quality gates and an access model that makes data usable without making it exposed.

  • Applied AI & agentic workflows

    Use-case shaping, model selection, retrieval over your own corpus, and agentic workflows with human-in-the-loop checkpoints and a complete audit trail.

  • AI governance & assurance

    A control plane for AI: model inventory, risk classification, evaluation gates, prompt and output logging, and the documentation your board and regulator will ask for.

How we engage

From mandate to handover, in four phases.

  1. Phase 01

    Discovery & insight

    Workshops with business and technical owners, data inventory, constraint mapping and a shortlist of use cases worth building.

  2. Phase 02

    Architecture & sizing

    Reference architecture, GPU and storage sizing, residency and control mapping, bill of materials and a costed sequence.

  3. Phase 03

    Build & integrate

    Deployment, data pipelines, identity and network integration, evaluation harness, and the first use case in production.

  4. Phase 04

    Handover & enablement

    As-built documentation, runbooks, enablement sessions and a governance model your own team operates.

What you receive

  • Reference architecture and annotated bill of materials
  • GPU, storage and network capacity model with growth assumptions
  • Data classification scheme and access control matrix
  • Deployment runbooks and as-built documentation
  • Model and use-case inventory with risk classification
  • Enablement sessions for your platform and data teams

How it maps to the frameworks

NCA ECC-2:2024
Asset management, identity and access, data protection and third-party control domains.
PDPL
Lawful basis, classification, minimisation and data-subject handling designed into the data layer.
SAMA CSF
Applied where the platform serves a financial institution or payment flow.
Vision 2030
Sovereign AI capability built, documented and operated in-Kingdom.

Why Next Step for AI

  • Infrastructure-first

    We come from datacenter and network engineering, so the platform underneath the model is designed by people who have built one.

  • Compliance in the design

    Residency, classification and evidence are architectural decisions here — not a remediation exercise after the audit.

  • Built to hand over

    Documentation and enablement are deliverables, not afterthoughts. You own the platform when we leave.

Start with an architecture, not a pilot.

Tell us the mandate and the data you hold. We will come back with a sovereign AI architecture, a sizing model and a costed sequence.

Book an executive briefing