Every organization thinks their data works.
Until it needs to.
Most systems are built for the good day. They pass acceptance testing, then something real arrives and the numbers stop agreeing with each other. I build for the bad day.
Built for
- NATO
- OECD
- Treasury Board of Canada
- ISED Canada
- Canadian Intellectual Property Office
The Pattern
The reporting cycle takes three months.
The model lives in a spreadsheet, and the one person who understands it has other work. The dashboard is accurate about last quarter.
By the time a decision reaches the executive floor, the number in the briefing has been through several hands since it left the source, and nobody in the room can name the last person who checked it. The analysis team spends most of its week cleaning data.
The technology is usually fine. The architecture underneath it was never designed for the question being asked, and nothing about that looks urgent until the week it becomes urgent.
What I build
- 01Interoperability specificationsFormal standards that define how systems from different organizations must behave when they meet. I write the requirement, implement it in code, then test a live system against the published document.
- 02Data platformsEnd-to-end systems from bare-metal infrastructure to front-end interface, engineered for restricted and highly regulated environments.
- 03Analytical pipelinesAutomated, version-controlled workflows that eliminate manual data preparation and produce reproducible, auditable output.
- 04AI integrationCustom agent pipelines and MCP servers that allow plain-language querying of secure, structured datasets.
- 05Institutional defenseThese systems get used by people who will be asked, in public, to explain them. I can defend a design decision to the team that built it and to the minister who signs the submission, using the same facts and two different vocabularies.
If any of that sounds like your organization, write to me and describe it. I will tell you whether it is something I can fix.
Tell me what is breakingFirst Principles
Six things that ten years inside regulated and politically exposed programs taught me. Each one has the full essay behind it.
01Failure-Mode ArchitectureAnything built only for the happy path is a liability, which is why the design starts at the shear plane.Read the essay 3 min- 02Hard BaselinesA denominator that moves will make any metric look like progress. I anchor to baselines that hold.3 min
- 03Architecture of AttentionThe interface that shows everything decides nothing. Mine shows what requires action and mutes the rest.3 min
- 04Consensus by SubtractionDozens of sovereign entities, no shared preference among them. What they hold in common is a list of things they cannot tolerate, and that list is the floor I build on.3 min
- 05Privacy-Safe AnalyticsPrivacy law takes the primary key away. Probability replaces it. High enough to act on, short of confirming an identity.3 min
- 06Technical SurvivalThe machine has no voice in a Cabinet room. I hand leadership the argument that keeps the capability funded.2 min
If you are building something that has to hold, write to me.
- Massachusetts Institute of TechnologyMicroMasters, Statistics and Data Science
- Carleton UniversityMaster of Applied Business Analytics
- Queen's UniversityBSc, Mining Engineering