Research

Across software, commodity trade, farming, fleets, law and sport, I kept arriving at the same problem from different directions.

For a long time I treated each project as its own problem with its own solution.

Looking back, they share a structure. Activity happens in the world; evidence about it is incomplete, informal, or held by different institutions; and someone has to decide how it becomes a record.

Building taught me two distinctions I now write about. First, integrity is not truth: a signed, timestamped record of GRADE = A proves who said it and when, not that the grade is right. Second, data is not observability: the same shipment can be fully visible to the warehouse holding it and almost invisible to the lender asked to finance it.

How should real-world activity become structured, trustworthy data when the underlying information is incomplete, informal, or distributed across institutions?

The systems were how I explored it. I keep working on it through five lenses:

Representation: what a record can and cannot capture about the activity it describes.

Verification: how confidence is established when evidence is partial.

Interoperability: how records survive moving between institutions.

Governance: who decides what counts, and who is accountable.

Uncertainty: how to represent what is unknown rather than hide it.

1. Research framework · 11 October 2026 · 18 minVerifiable State in Multi-Agent Freight NetworksThe joint design of observation and decision thresholds in adversarial trade networks

   A signed, timestamped record proves who asserted something and when, not that it happened. The paper asks where verification should sit in a freight network, how confidence should be estimated from unreliable evidence, whether settlement can reward truthful reporting, and whether verifiability changes the cost of trade finance. Its central proposition: observation and decision thresholds must be designed together.

2. Research concept paper · 11 October 2026 · 7 minInformation-Aware Resilience in Multi-Agent Shipping and Supply NetworksHow information quality changes disruption propagation, adaptation and recovery

   Proposes that information quality is itself a determinant of resilience. An agent-based model runs the same disruption under perfect, fragmented, delayed, noisy and shared-state information, and treats where to observe as an optimisation problem: minimise C(S) + λL(S), the cost of observation against expected loss.

3. Essay · 11 October 2026 · 9 minThe ESG Observability GapData as shared infrastructure for climate resilience and sustainable finance in commodity markets

   Defines the gap between a supply chain's physical state and the representation an institution can act on, O = D(X, X̂), and argues that the same evidence serves both climate resilience and sustainable finance. The claim is conditional: where the cost of proof falls on small producers, better observability can deepen exclusion.

4. Analysis · October 2026 · 14 minThe Next Financial Infrastructure for Global Trade Will Be Built on StablecoinsFrom commodity settlement to institutional treasury, OTC liquidity and trade finance

   Argues that the larger opportunity for stablecoins in trade is not faster transfers but connecting treasury, institutional liquidity, settlement and trade finance around the commercial transaction underneath them.

AXK Network: A Verification-and-Settlement Layer for Cross-Border Commodity Trade in East Africa. ALU undergraduate capstone, 2026: a 21-page study with literature review, methodology and five system diagrams.

AXK Network White Paper, v1 published; v2 forthcoming.

Treasury: Training an AI Co-founder Agent (2026). Research paper and working agent for opportunity search, fundraising workflow and correspondence, with human approval gates.

Open-Source Components for Digital Traceability and Data Mobility. In preparation.

Research interests: institutional evidence and governance; digital public infrastructure for informal and cross-border trade; portable legal and personal records; the social consequences of measurement and verification.

Building showed me where the problem is. The work now is understanding it properly.

This is the question itself.