Deanna Gallimore

AI Systems Builder | Decision-Support Architect

I build decision-support AI systems that earn trust before they ask for it — agentic workflows, decision engines, and gamified learning platforms grounded in explainability, governance, and human-in-the-loop design. Currently shipping at LinkedIn.

(§) Production Pipeline (◉) Active R&D Lab
deanna.gallimore@proton.me | LinkedIn | Resume (PDF)

How I Build AI Systems

1. Explainability first
2. Confidence scoring with human-in-the-loop
3. Override and feedback as core mechanics
4. Evaluation as a first-class feature
5. No black boxes

Production Pipeline (§)

SignalForge GTM

Production Pipeline (§)

Bottleneck: Hidden risk in pipeline data and lack of AI auditability.

Explainable agentic pipeline for CRM/Call reconciliation with full source-attribution and human-in-the-loop governance.

Python | Agentic Workflow | GTM Ops

Breadcrumbs

Production Pipeline (§)

Bottleneck: Non-technical employees have ideas but no guided path to build them within enterprise-approved tools.

A guided builder concierge that walks any employee from "I have an idea" to "I have a working build" — using only approved tools, with governance baked in at every step.

Copilot Chat | Declarative Agent | Decision Tree | LLM Artifact Gen
Dossier →

Enablement Command Center

Production Pipeline (§)

Bottleneck: Enablement work scattered across plans, metrics, artifacts, and status narratives.

Enterprise enablement command center unifying program lifecycle management, collaborative tasks, KPIs, and executive narrative generation with privacy-conscious people display.

Power Apps | React | TypeScript | TanStack Query | Dataverse
Dossier →

Training Tavern

Production Pipeline (§)

Bottleneck: Low engagement and no measurable skill development in enterprise training.

A 5-game AI-powered RPG training platform with branching dialogue, companion unlocks, Kirkpatrick L1-L4 reporting, a document-to-scenario pipeline, and 27 custom assets. Built in Superblocks, designed for portability.

React 19 | TypeScript | 5 AI APIs | WCAG 2.1 AA | Scribe's Desk
Dossier →

Escalation Decision Engine

Production Pipeline (§)

Bottleneck: Escalation ambiguity and inconsistent triage decisions.

Weighted decision engine integrating knowledge retrieval, historical pattern analysis, and risk scoring into a unified workspace with dual-mode interface and closed-loop feedback.

Python | Streamlit | Decision Logic | Explainable AI

Quickfire Quadrant Quest

Production Pipeline (§)

Bottleneck: Organizational prioritization blindness and siloed enablement workflows.

Multi-tenant decision-matrix platform tracking Kirkpatrick-aligned analytics to map behavior change to business outcomes.

PostgreSQL | Drizzle | Express | React

Trust Vector Arcade

Production Pipeline (§)

Bottleneck: Human error in high-pressure triage.

Gamified simulation engine using L1-L4 telemetry to quantify decision accuracy.

Netlify | JS | Telemetry

Audit-Gate Pipeline

Production Pipeline (§)

Bottleneck: Non-deterministic AI outputs.

CI/CD-integrated adversarial pipeline enforcing safety logic-gates.

Python | Streamlit | Git
Dossier →

Active R&D Lab (◉)

Devotional Intelligence Suite

Active R&D Lab (◉)

Bottleneck: Stagnant, passive spiritual content usage.

A Gamified Discipleship Engine that treats spiritual growth as a high-stakes quest. Orchestrates RAG-based biblical application.

React / Vite | Supabase | LangChain.js
Dossier →

Silas-Creed Architecture

Active R&D Lab (◉)

Bottleneck: Narrative state-drift in 500k+ word IP.

Relational state-machine tracking plot-critical flags for canonical consistency.

Airtable | Git | VS Code
Dossier →

Remi: The Refiner's Quest

Active R&D Lab (◉)

Bottleneck: Need for high-integrity formation.

Modular RPG engine using recursive gameLoops and Euclidean state-management.

JS | Canvas API | GitHub
Dossier →