Case studies

Platforms, boundaries and AI systems in production.

Selected work from my CTO role, earlier infrastructure roles and fractional CTO engagements. Each one names the constraint, the decisions and what changed.

Multi-cloud Kubernetes platform · 2025 – present

Ten minutes to fifty seconds: shared control planes for tenant provisioning

Redesigning tenant provisioning on a multi-cloud Kubernetes platform, from a dedicated VM per customer to shared hosted control planes.

~50 s tenant provisioning, down from 8–10 minutes

Messaging platform (CPaaS) · 2024 – 2025

Forty apps, twenty databases, ten teams, zero downtime

Moving a messaging platform that handles more than 10 billion messages a year from AWS EKS to Oracle Cloud OKE without customers noticing.

$15–20K/mo lower cloud spend, 10–15% of the monthly bill

Travel · 2025 – present

An internal assistant that can admit uncertainty

A retrieval-based knowledge assistant for a travel company of about 1,300 people, built against a board-demo deadline and taken to production.

~1,300 employees who can query company policy in plain language

Healthcare software · 2025 – present

From 22 pentest findings to an audit-ready position

Turning a stalled enterprise security review for a healthcare SaaS into a SOC 2 readiness program with evidence behind it.

22 penetration-test findings remediated

Manufacturing · 2025 – present

Get personal data under control

Practical GDPR remediation for a manufacturing intelligence platform: data lifecycle audit, deletion, retention and process.

Media · 2025 – present

Make cloud and GPU spend visible

Steady AWS and EKS operations for a media company, including GPU node provisioning and closer attention to how that capacity is used.

Ten minutes to fifty seconds: shared control planes for tenant provisioning ↗

Forty apps, twenty databases, ten teams, zero downtime ↗

An internal assistant that can admit uncertainty ↗

From 22 pentest findings to an audit-ready position ↗