Speaker and builder
Elevatech Summit London 2026
8 August 2026 · London, UK
Artificial Intelligence (AI) for Professionals
- 16 speakers
- 14 sessions
- Built Connection Compass live
Senior Data & AI Leader · London, UK
I lead strategy, architecture, engineering and delivery for AI in regulated, industrial and other high-stakes environments.
13+ years
across engineering, consulting, data and AI
Chevron → Accenture → Capco
industrial engineering to AI leadership
National public-health delivery
led a 20-person AI innovation team
AIGP
Azure AI Engineer, UCL, Imperial and Royal Holloway
Operating model
01
Identify worthwhile opportunities, define value, and make ownership explicit.
02
Connect data, architecture, engineering, product, teams and delivery.
03
Establish evaluation, security, governance, adoption and reliable performance.
Selected work
Architecture and engineering standards for a national public-health surveillance platform designed to absorb a tenfold surge without losing trust in the data.
250k records/dayNational public healthA service designed around the people who need public-health guidance, not around the organisational chart that produced it.
Alpha → BetaConsumer goodsAssurance that changes a deployment condition is worth more than assurance that produces a longer document.
Deployment conditions changed
Speaking into building
At Elevatech Summit London in August 2026, a 30-minute breakout becameElevatech Connection Compass: a working product that matched attendees on-device and explained every recommendation.
Career throughline
01
Learned to treat reliability, maintenance and safety as design constraints, not afterthoughts.
02
Carried that operating discipline into complex programmes, architecture and delivery.
03
Connected strategy to engineering, product and adoption across regulated environments.
04
Now focuses on the evidence, controls and human judgement that make AI dependable in use.
Point of view
Note 01
A model call is one component. The system around it carries the data, controls, evaluation and ownership.
Note 02
Component-level assurance is not enough when authority, context and failure modes cross tool boundaries.
Note 03
A human in the loop is not a control unless that person can see enough, act in time and change the outcome.
Speaking and teaching
Speaker and builder
8 August 2026 · London, UK
Artificial Intelligence (AI) for Professionals
Guest lecturer
Recurring · London, UK
Predictive, prescriptive and risk analytics
Panel speaker
24 May 2024 · Rhodes House, Oxford
AI for Inclusive Development in Africa
Next conversation
I’m targeting Head of Data & AI, Director-level Data & AI, AI platform or product leadership, and advisory work on dependable enterprise AI.