Case Studies
Client results
Organisations bring me in when they need senior expertise to get AI and cloud platforms into production. Here's a selection.
HSBC — ML Platform Migration
Lead Cloud / Platform Architect
Oct 2025 – Present · London, UKChallenge
Migrate a self-managed Kubeflow-based ML platform to GCP Vertex AI within a highly regulated banking environment with strict security controls, no console access, and SOX-compliant workflows.
Approach
Designed secure platform services operating within VPC Service Controls with restricted egress. Defined workload identity patterns using GCP Workload Identity Federation. Built CI/CD pipelines with SOX-compliant approval workflows, container security scanning, and controlled artefact promotion.
Outcome
Delivered an enterprise-grade ML platform architecture that satisfied both engineering and regulatory requirements. Established a secure software supply chain with private artefact repositories and approved base images.
Carrier Corporation — Connected Devices & MLOps Platform
Cloud Lead Technical Architect
Feb 2022 – Sep 2025 · Palm Beach Gardens, FLChallenge
Build a global connected device platform for residential HVAC connectivity and an MLOps platform for predictive analytics, while also advising on post-merger technical consolidation following the Viessmann acquisition.
Approach
Designed and developed a Kubeflow-based MLOps platform. Advised on a Vertex AI ML platform for energy forecasting in partnership with Google Consulting Services. Led multiple distributed product teams using AWS cloud services with serverless and containerisation patterns.
Outcome
Delivered a production platform connecting residential HVAC devices globally. Established DevOps practices across multiple teams and provided technical guidance for the Viessmann Azure platform integration.
EY Global Innovation — ML at Scale
Cloud Technical Architect / DevOps
Aug 2018 – Apr 2019 & Apr 2019 – Oct 2019 · San Jose & LondonChallenge
Design and deliver a globally scalable platform for machine learning at scale, then lead a major AWS-to-Azure migration across the organisation.
Approach
Built a Kubernetes-based platform on Azure with Databricks and Apache Spark. Implemented GitOps workflows with FluxCD. Designed DevOps processes enabling data science and engineering teams to collaborate on ML delivery.
Outcome
Delivered a production ML platform serving EY's global innovation teams. Successfully migrated workloads from AWS (S3, EC2) to Azure Kubernetes Service with zero downtime.
Roche — Clinical Platform Rescue
Azure Technical Architect / DevOps Lead
Aug 2021 – Dec 2021 · Basel, SwitzerlandChallenge
An in-house clinical study management platform was being migrated to Azure. Microsoft Consulting Services had created the initial PoC, but it had become overly complex and the team was struggling to get it production-ready.
Approach
Rationalised and streamlined the CI/CD approach in Azure DevOps. Migrated from Azure ARM templates to Terraform to simplify complex infrastructure management. Standardised Kubernetes deployments with Helm and Kustomize.
Outcome
Rescued a stalling project and delivered a production-ready clinical platform. Simplified the infrastructure to a point where the in-house team could maintain and extend it independently.
Caterpillar — Global Service Platform
Azure Technical Architect / DevOps Specialist
Aug 2018 – Dec 2023 · Chicago, ILChallenge
Design and deliver a new global service platform supporting Caterpillar's service and maintenance partners worldwide, with multi-region deployment requirements.
Approach
Architected a Java Spring Boot microservices platform leveraging Azure Web Applications, Traffic Manager, Application Insights, KeyVault, and SQL. Designed a CI/CD pipeline in Azure DevOps for continuous testing and multi-region deployment.
Outcome
Delivered a globally distributed service platform with automated multi-region deployments. The platform supported Caterpillar's worldwide partner network with high availability and consistent performance.
Smartlogic Semaphore — Multi-Tenant SaaS Platform
Cloud Technical Architect / DevOps
Dec 2014 – Feb 2018 · London, UKChallenge
Design and build a multi-regional, multi-tenant cloud platform for document classification, processing millions of documents per day for some of the world's largest organisations.
Approach
Architected a microservices platform on Azure using Service Fabric, Azure AD, Key Vault, OMS, and Application Insights. Built CI/CD processes with VSTS and Jenkins. Later developed Office 365 and SharePoint integrations using React, Redux, and Azure services.
Outcome
Delivered a production SaaS platform classifying millions of documents daily for global enterprise clients. The platform ran reliably across multiple Azure regions with full multi-tenancy.
Have a similar challenge?
I've delivered cloud platforms across banking, pharma, industrial, media, and enterprise SaaS. Let's discuss how I can help your organisation.