Services

Agentic development

I build production software hands-on with AI coding agents. The agents write a lot of the code; I do the parts that decide whether a project succeeds. I frame the problem with you, choose the architecture, break the work into steps an agent can carry out, and review and test every change before it ships. A career that began in enterprise support and continued through software testing taught me to investigate problems and explain the next step clearly.

Projects start with a written specification: what we’re building, who it’s for, and how we’ll know it works. From there I work the way a disciplined engineering team does: small changes reviewed in version control, tests and checks before anything ships, security and accessibility treated as requirements rather than afterthoughts, and documentation the next developer can actually use. You get the speed of AI with software you can maintain, extend and trust. I’ve packaged these habits into an open-source skill that any AI coding agent can follow, and written up how I build with agents, using this website as the example.

Confidential AI

Some information shouldn’t leave your organization: client files, contracts, financial records. I set up open AI models, such as Qwen and Gemma, to run on machines you control, so your team can summarise confidential documents and analyse financial data without sending any of it to an outside AI service.

You get the useful parts of AI, with your data staying where it is, alongside a clear AI-use policy so everyone knows what’s allowed.

Technical project management

I’ve led testing, projects and releases for Fortune 500 companies, often with teams split across countries and time zones. I turn a vague goal into a plan with a clear scope, owners and dates, keep dependencies and risks visible, and tell stakeholders early when something needs a decision. Regular updates show what is done, what is uncertain and what needs attention.

Whether you need someone to run a project end to end or to steady one that’s drifting, I plan for the risks up front, keep the work moving in small steps you can see, and stay calm when plans change.

Solutions architecture

Good architecture starts with the business, not the diagram. I’ll help you decide what to build, what to buy and what to leave alone. When speed matters, I reach for proven tools that ship in days and put custom work only where it makes a difference.

The aim is a design that copes when demand jumps overnight, and that your team can understand, run and change, with or without me. You get full access to the code, alongside documentation that helps another developer take over.

Quality, security and infrastructure

My years in support and testing still shape how I build and operate software. I run regression, performance and security testing, and I look after security and infrastructure for the businesses I work with: access control and multi-factor authentication, email authentication with DMARC, DKIM and SPF, domain and certificate management, server administration and firewall rules.

Fractional CTO

Many of my clients don’t have a technical owner, and don’t need a full-time one. I fill that gap part-time: keeping systems running and secure, planning technical work across the year, and writing the policies that let your team use AI safely. We usually start with a short assessment and a written plan, then build what matters most, and then I stay on to keep it running.

Since August 2026 I’ve been the fractional CTO for Bedrock Sandals, after 14 years looking after its security and infrastructure. The work covers projects, QA, troubleshooting, policy and security, infrastructure, automation and AI, coordination and day-to-day operational support.