About

About Me

I've been fortunate to build a lot of interesting systems over the past thirty years — intelligent agent compilers in the 1990s, 3D engines and plant-simulation software, Machine Acts (my PhD work on giving machines a memory), the Au programming language, and, today, secure agent systems at Nearfield.ai. What ties them together is a foundation in computer science: compilers, declarative languages, and machine memory.

At Nearfield.ai I put that experience to work building agent systems enterprises can actually run in production — solving genuinely hard problems while keeping everything on secure, local infrastructure, where data sensitivity, accuracy, and control really matter.

In the 1990s I worked in research and development at ICI, an industrial partner on the ADEPT project (Advanced Decision Environment for Process Tasks) — a multi-agent system for managing business processes with negotiating, autonomous agents. I built ICI's agent demonstrator using Viper, an agent compiler and language I designed for creating intelligent agents distributed over a network.

At Nearfield.ai the work focuses on two areas. PIIQ automates Data Subject Access Requests (DSARs) and related compliance using locally deployed AI models — turning a multi-day manual process into hours while ensuring sensitive data never leaves the organisation. Beneath it sits a proprietary multi-agent infrastructure: OS/A (an Operating System for Agency), the Contextualiser (episodic machine memory), Agent Markdown, and Scope.

Scope is a new class of development environment designed to safely manage, structure, and constrain code generated by intelligent agents in production. Nearfield is now preparing to open-source core components of this infrastructure as it moves toward a broader developer platform. My PhD work on Machine Acts — capturing context for personal search — feeds directly into the Contextualiser's approach to episodic machine memory.

I am equally happy running a technical business as CEO or leading technology as CTO. I think differently — there is no problem that cannot be overcome, and I have little patience for "impossible". That can-do attitude is reflected in the systems I have built over my career, several of which were deemed the task of a lifetime. I lead from the front: if there's a deadline I'm the first online and the last to leave, and I'd never ask my team to do something I'm not prepared to do myself. I care deeply about nurturing talent — the developers who've worked with me would say they came out the other side better at their craft.

Research Areas

  • Intelligent Agent Systems & Orchestration
  • Compiler and Language Theory
  • Declarative Languages & Machine Memory
  • Knowledge Management & Ontological Modelling

Education

PhD, Computer Science — Loughborough University (1998–2004)

Thesis: Machine Acts: Capturing Context for Personal Search, supervised by Professor J. L. Alty; external examiner Professor Nick Jennings.

Experience Overview

CEO of Nearfield.ai, building secure enterprise agent systems; Co-CEO and CTO of Aucoda and inventor of the Au programming language. Over 30 years writing software, leading technology, and managing developers to build complex systems.