Case studies

Real infrastructure problems. Real constraints.

Each case study covers the environment, my role, what was designed and built, and the outcome, from today's sovereign cloud and AI engagements at SITE back to the bank-scale platforms where it started.

SITE · 2025 – Present

AI Adoption at SITE Cloud

Helping regulated enterprises adopt SITE's sovereign AI platform: scoping use cases, building RAG pipelines on in-Kingdom models, and taking teams from proof-of-value to production without data leaving the country.

In-Kingdom Data residency
PoV → Prod Adoption path
AI Adoption Sovereign AI RAG Regulated Industries

HUMAIN · Sep 2026 — Independent benchmark

HUMAIN m3 Inference Performance Assessment

I had the honor of testing HUMAIN inference serving: 16,126 requests against humain-m3-research-preview on the HUMAIN Node gateway. 21.6 req/s sustained, 11.5k output tok/s peak, sub-second first token — full report with scaling curves, tail-latency analysis, and production guidance.

21.6 req/s Sustained throughput
11.5k tok/s Peak output rate
7/10 Production readiness
HUMAIN LLM Inference Benchmark Sovereign AI

Alrajhi Bank · Nov 2021 – Aug 2024

VMware Platform at Scale: 300+ Hosts, 4,000+ VMs

Design, implementation, and operations of a multi-site private cloud spanning vSphere, NSX-T network virtualisation, VMware Cloud Director multi-tenancy, and SRM-based disaster recovery, serving one of Saudi Arabia's largest financial institutions and its subsidiaries.

300+ ESXi hosts
4,000+ Virtual machines
99.9% Availability
vSphere 7/8 NSX-T VMware Cloud Director SRM VxRail vSAN

Banque Misr · Sep 2024 – Feb 2025

OCI Cloud Operations & FinOps

End-to-end OCI ownership for one of Egypt's largest state-owned banks: platform implementation, network architecture, IAM hardening, day-to-day operations, and cloud cost governance built from scratch.

100% OCI platform ownership
Zero Knowledge gap at handover
OCI Cloud Operations FinOps Cloud Networking

Related reading: What running NSX-T across 4,000 VMs taught me about micro-segmentation · HUMAIN m3 inference performance assessment.

Working on something similar?

I'm always up for a technical conversation about cloud architecture, migration, or sovereign AI.