portfolio --selected-work
Selected work
AI, data & research
I build AI systems and lead the work that brings them into use. These projects span professional programmes, personal experiments and open source: what I contributed, the decisions I made, and the evidence behind the results.
7 projects / Research · Engineering · Leadership
- 01
ProfessionalFeatured projectIn production
A multi-agent research system that scores organisational AI maturity from public evidence alone, with published FTSE AIM 100 and FTSE 350 studies.

- Companies scored, FTSE 350 study (July 2026)
- 350
- Distinct public sources cited, FTSE 350
- 5,878
- Criteria across five pillars
- 180+
- Median research time per company, FTSE 350
- ~15 min
Read the case study → - 02
PersonalLive, free to use
A free web app for readers who have abandoned a book mid-way: upload the PDF or EPUB, mark where you stopped, and get a recap of the story so far with nothing from beyond that point.

- 03
ProfessionalDelivered August 2026
A discourse-pressure map of Israel–Palestine conflict language: 300+ client-approved terms scored for how reliably they provoke hostile reaction, each with a confidence rating, feeding a peace campaign's messaging.

- 04
ProfessionalLive, living research
An open audience-intelligence matrix profiling ten UK water-user communities across seven strategic dimensions, so an antimicrobial resistance campaign can reach people who have never heard of the threat.

- 05
PersonalOngoing; run 34 published August 2026
An evidence pipeline and journalistic report over eight decades of declassified US UAP records, scoring each artefact for documentary strength rather than for the likelihood of anything exotic.

- 06
HackathonHackathon entry, July 2025
A real-time audio commentator for multi-agent AI workflows, built for the Google Gemma 3n Impact Challenge: sensitive agent decisions run locally on Gemma 3n while a cloud commentator explains them to a human audience as they happen.

- 07
Open sourceMaintained since March 2025
An open-source deployment guide for running the n8n automation platform serverlessly on Google Cloud for a few pounds a month, with more than 600 GitHub stars and contributions from Google's Cloud Run product team.
