Regis Chapman
Summary
I build the layer that decides whether an organization’s information can be trusted, by the people reading it and by the AI systems reading it. At Octant Labs I owned the knowledge layer of a funding program through seven consecutive rounds: the intake schemas, the evaluation data models, the validation workflows, and the dashboards leadership planned from. Alongside that work I authored and published a body of open standards specifying what makes a claim checkable, and what a system owes a reader about the material its answers rest on.
Those standards are not proposals. They run: a versioned corpus in daily use, validation that fires at write time, a drift sweep across every published surface, and three MCP servers that make the standards callable inside any AI assistant. Underneath is three decades in systems and network architecture at Apple, Sun Microsystems, and Microsoft, and eight years as sole senior administrator for a government of 500+ users. I read how a system is actually built, find where its parts have stopped agreeing, and make it run.
Selected Work
Published Standards & Working Systems
- ORE, a specification for grading data sources at the ingestion boundary (CC0). It grades sources as uncertainty rather than as quality verdicts, forbids party count from standing in for independence, requires every output to expose the reliability profile of the material it rests on, and names three intake postures with enforceable downstream conditions, including a quarantine boundary for bulk ungraded material that may serve search but may not support a decision.
- CROSS, WALKRI, GRAIN, and the CRAFT meta-standard (CC0). WALKRI holds any schema field to the instrument test: working from the definition alone, would two independent readers collect the same thing. CRAFT requires that conformance be assessable from published declarations alone, without access to any implementation, which is what makes a standard adoptable across teams nobody can audit.
- The Coordination Structural Integrity Suite, ten standards covering precision-first design, information asymmetry classification, structural consent legibility, and sensemaking capacity. Adopted by Regen Coordination as the structural integrity framework for the Regen Web3 Toolkit, registered as a source in its public data layer, and mapped against their architecture in a published alignment report.
- A working context engine, built and operated for two years: a versioned corpus where every document declares whether it is canonical or superseded and names its successor, validation that fires automatically on write, a drift sweep across every published surface and built artifact, and a resource registry mapping remembered names to deployed titles to platform slugs.
Practice
I build AI-native and verify mechanically. The tools treat model output as a claim to be checked rather than an answer to be trusted: evaluation against criteria stated before the material is seen, contrast ratios computed rather than asserted, findings that cite and quote the evidence they rest on. What I add to a codebase or a corpus is the part that says how a claim gets checked and who is allowed to rely on it.
Skills
Experience
- Owned the operation’s knowledge layer: the intake schemas, evaluation data models, and grantee records that made a distributed program legible, replacing scattered documentation with one structured source of truth the team worked from daily.
- Designed the taxonomies, information architecture, and validation workflows that kept that record accurate as it grew, so what the system reported and what was actually true did not drift apart across seven rounds.
- Built the reporting layer on top of it: performance dashboards on Vercel and evaluation pipelines in Linear, Notion, and Airtable that leadership used for planning and operational review.
- Owned seven consecutive funding rounds from initiation to delivery, an operating cadence spanning intake, evaluation, disbursement, and reporting, including a full platform migration I scoped and delivered while rounds were live, carrying 159 project grants across it without operational continuity breaking.
- Evaluated nearly 800 applications individually across the rounds, more than 450 in a single round, and published the standards the operation ran on as open specifications.
- Co-founded and led a conflict-resolution organization; built dispute-resolution frameworks adopted by 15+ communities with a 90% resolution rate, and trained 120+ members across 8 cohorts.
- Advised organizations on coordination and organizational development; authored original frameworks, including The Virtuous Cycles and a decolonization framework.
Earlier
- Senior IT System Administrator (titled Information Technology Technician), Cowichan Tribes, a First Nations government (Nov 2016 to Nov 2024): infrastructure for 500+ users and 150+ virtual machines; server, virtualization (VMware/ESXi), networking, and VoIP.
- Senior Ecommerce Consultant, Grow Mushrooms Canada (Nov 2019 to Jan 2025): freelance, part-time alongside the full-time roles above; system administration, site architecture, and marketing automation for an ecommerce business.
- Non-Windows DevOps Lab Engineer, Microsoft (contract, 2011 to 2014), and System Architect, UniversalMac (2014).
- IT Director, Graphic and Web Designer, and Head Yoga Teacher, Sivananda Ashram Yoga Farm (2005 to 2008); Phone Support Engineer, Apple (2002 to 2005) and Sun Microsystems (1998 to 2000); earlier technical roles 1994 to 1996.
- Founder and Head Instructor, Silent Motion Yoga (Mar 2009 to Aug 2024).
Education & Certifications
Sivananda Ashram Yoga Farm (2005)
Links
This page carries my public contact details. For a copy with full contact information, get in touch through the site.