Projects
Professional and independent systems I have designed, built or helped take into operation.
A progressive reference design for an enterprise that wants to move from isolated AI pilots to governed delivery at scale without creating a new portal, platform or integration for every use case.
- Standardise shared paths only after real delivery begins to repeat.
- Make evaluation, identity and governance part of the delivery path.
- Measure cost per successful business outcome, not per token.
Ongoing - developed progressively since 2024
Project Description
I co-authored the group’s 2026 AI strategy, now in execution as more than 70 initiatives across 19 divisions, together with the governance model it depends on - data residency, model approval and release gating, aligned to the NIST AI Risk Management Framework and ISO/IEC 42001.
Key insights
- Give every initiative a named owner and a measurable outcome.
- Turn repeatable governance claims into executable delivery gates.
- Keep evidence, adoption and cost visible after launch.
Duration taken to do the project
Multi-year programme - ongoing
Project Description
A self-hosted personal data platform and governed agent mesh. Around 590,000 bi-temporal observations from 22 connectors land in a single fact shape, get answered by local models, and cannot leave the machine without a per-action approval.
Key insights
- Normalize every personal source into one bi-temporal fact shape.
- Enforce privacy through capability isolation and per-action approval.
- Keep inference local by default and fail closed when cloud access is absent.
Duration taken to do the project
Ongoing - started in 2026
Project Description
A commercial, multi-tenant SaaS for road and utility cross-section design. Civil engineers compose a street cross-section and the utility corridor beneath it, then issue it as PDF, PNG or CAD. I built and operate all of it - product, tenancy, billing, security, performance and deployment.
Key insights
- Prove tenant isolation with an attack harness, not a policy claim.
- Optimize measured bottlenecks before adding infrastructure.
- Treat billing, security and operation as part of the product.
Duration taken to do the project
Ongoing - started in Mar 2026 and launched in Jul 2026
Project Description
I built the group’s executive analytics platform from scratch - 27 real-time dashboards opened weekly by senior management, with daily executive notifications, KPI-threshold alerting and an embedded chat layer. It became the data foundation the AI platform was later built on.
Key insights
- Shared KPI definitions create more value than another dashboard.
- Governed data and reusable metrics serve both analytics and AI.
- Run integration and data engineering as measured products, not ticket queues.
Duration taken to do the project
Ongoing evolution - started in 2024
