Skip to content

Applied AI Systems

Systems that survive contact with reality.

Selected systems and case studies showing the problem, the architecture, what was actually built and how it was checked. Models reason; authoritative systems settle the claim.

Every case in one order: Problem → Architecture → What was built → Verification → Evidence → Status

In Development2025 – present

Healthcare policy intelligence: which revision applied, and when

An architecture for policy decisions where an apparently correct answer can still be wrong because it rests on the wrong revision, date or authority: a 2025 prototype that resolves the revision in force on the date, and a proposed modernization that reconstructs the source first and reasons over procedure behind deterministic gates.

Senior AI Architect (independent project) · DeNovo Innovations Group

In productionJan 2026 – present

A production generative-AI platform, built with an evidence-gated human-AI engineering system

Built end to end by a human-AI hybrid engineering system.

AI authored and reviewed; a human directed, verified and accepted, using the DeNovo human-AI hybrid engineering toolkit. A production generative-AI platform and the applied research environment for scalable AI systems, model evaluation, provenance and operational reliability.

Chief AI Officer / AI Architect (engagement via DeNovo) · AllMyMascots, LLC

The DeNovo toolkitThe human-AI engineering system that built AllMyMascots: agents with bounded roles, controls that must be able to fail, and an evidence layer over everything published.
Diagram of the DeNovo human-AI engineering system: human direction, role-separated agents, repository context, verifiers, hooks and retained evidence; solid borders mark implemented parts and dashed borders planned ones.
The DeNovo toolkit. Solid border = implemented, dashed = planned.View full toolkit diagram ↗See it on Evidence & Provenance →
Why the controls exist →

A production generative-AI platform, built end to end by a human-AI hybrid engineering system

The platform is AllMyMascots. AI authored and reviewed; a human directed, verified and accepted. The concept architecture and an end-to-end trace, with the repository-verified diagram beneath.

AllMyMascots · concept architecture

VERIFIED — PUBLIC
Illustrated concept architecture of the AllMyMascots generative-AI platform, April 2026: client, application, workers, model providers, storage and planned components.
Concept architecture, April 2026. Payments, fulfillment, S3/Glacier tiering and the graph store are planned components. The four-stage DPO, RLHF and data-collection pipeline is proprietary to AllMyMascots and held as trade secrets.sha256 54cb3a2b8a5ca06a… · the author's own figure

AllMyMascots · end-to-end trace

VERIFIED — PUBLIC
Illustration of an end-to-end request trace through the platform with a provider fallback, showing spans across web, bridge, queue, workers and model providers.
End-to-end trace.sha256 9e9037eacecc0441… · the author's own figure

Verified architecture: application architecture and human-AI engineering system — repository-verified diagram, snapshot 85524df, reviewed 2026-09-06.