Human purpose.
Evidence. Governed AI.
I build AI-enabled products and workflows that help people examine evidence, understand complexity, and finish important work. The technology expands what one senior person can analyze and produce. I remain accountable for meaning and quality.
Built, not merely discussed.
Live products and working concepts, each one built with the Clarity System below. Status is stated on every card, including what is still a prototype.
Live platformNILOMETER
A source-grounded platform connecting resilience risk, development finance, economic exposure, and human stories across more than 200 economies. Its AI analyst answers from available country evidence and exposes the claims behind its facts.
Visit thenilometer.org ↗
Working prototype · pilotPublication Readiness Audit
A local, privacy-first audit that helps teams complete a structured final review before an institutional Word report publishes. Nothing is uploaded, stored, or changed. Read how it was built, what it got wrong first, and how that was fixed.
Read the case study →
Independent conceptResilience Digital Experience
An independent, non-official design concept exploring how human stories, institutional priorities, impact metrics, publications, and global reach can become one coherent digital experience.
View the concept ↗Staff, not authorities. Each system is assigned a job and kept to it. None of them is the evidence base, and none of them makes the final call.
- GeminiMultimodal analysis and large-scale source review.
- ClaudeDeep reading, structured reasoning, and editorial development.
- OpenAIResearch, data analysis, creative production, and workflow automation.
- GrokReal-time public-discourse monitoring. Discovery only, never the source for an institutional claim.
Confidential client material is used only in business or API environments with documented retention and privacy controls, never in consumer accounts.
From human purpose to accountable delivery.
Seven steps, in this order, on every engagement. AI works in the middle. People own both ends.
AI accelerates the work. Evidence grounds it. Human judgment governs it.
Who has to decide what, and what changes for them if the work succeeds.
Approved data, documents, interviews, and context. Claims start here or they do not start.
Models compare, question, and organize the material and produce options. Working material, not a verdict.
Facts, sources, calculations, quotations, and contradictions checked against the approved evidence.
Subject knowledge, cultural awareness, editorial judgment, and the politics of the room.
Findings become narratives, reports, presentations, visualizations, and digital experiences, then get tested on real readers.
Final human review. Delivered with visible sourcing and stated limits.