Independent Applied-AI & Workflow Builder

I build local-first software, deterministic AI-assisted workflows, and technical-operations tools — then document what works, what fails, and what outside reviewers challenge. My work sits at the intersection of local-first software, human-in-the-loop AI, workflow design, systems thinking, technical operations, and practical product development. I prioritize systems that are inspectable, testable, useful, and honest about their limits.

Portfolio Hub · LinkedIn · Email

These are the fastest ways to see what I actually build. Each project emphasizes distinct technical themes, live demonstrations, and transparent architectural decisions.

A local-first customer-evidence and product-decision workspace.

• Technical Themes: Evidence organization, claim verification, structural integrity.
• Live Demo: Try Validation Ledger
• Feedback: Challenge the evidence model

Local-first visual workflow canvas with portable projects, snapshots, browser validation, and consent-gated AI assistance.

• Technical Themes: Canvas/SVG, local-first storage (localStorage), end-to-end productization.
• Live Demo: Try WeaveStudio
• Feedback: Test WeaveStudio and provide feedback

Deterministic systems simulation with typed graph models, cascade analysis, reproducible reports, and an original stability heuristic.

• Technical Themes: Typed graph models, deterministic simulation engines, cascade analysis.
• Live Demo: Try BuildWorld AI
• Case Study: Technical decision case study

A commercial white-label quote-calculator package for agencies and local-service web implementers.

• Technical Themes: Typed quote logic, licensing, embeddable delivery, WordPress integration.
• Live Demo: Try QuoteForge Local

(Current public work also includes ProcessHarbor and ScamShield AI.)

• Problem Framing: Identifies users, workflows, constraints, non-goals, and success criteria before implementation.
• Systems Thinking: Models dependencies, bottlenecks, state transitions, failure modes, reversibility, and evidence boundaries.
• Applied AI Judgment: Separates deterministic core logic from optional provider assistance and requires human review before generated work is applied.
• Technical Operations: Produces SOPs, checklists, knowledge-base drafts, gap reports, audit records, version histories, exports, and operational documentation.
• Product Execution: Moves from concept through implementation, testing, live review, documentation, packaging, and transfer guidance.
• Claim Discipline: Distinguishes implemented, tested, deployed, experimental, legacy, and unverified capabilities.

Technologies: React · TypeScript · Vite · Next.js · Zod · Vitest · Playwright · Canvas/SVG · localStorage / IndexedDB
Methods: Deterministic workflow design · Structured exports · CI verification · Technical documentation · Human-in-the-loop AI boundaries

This is an AI-assisted portfolio. I direct product strategy, requirements, workflows, scope boundaries, acceptance criteria, verification expectations, source authority, and public claims. AI tools assist with implementation, research, debugging, testing, and drafting; I inspect, revise, reject, validate, and package the resulting work.

Review the supporting evidence:

• Portfolio evidence dossier — Capability-to-evidence matrix and reviewer guidance.
• Technical decisions — Architecture choices, tradeoffs, alternatives, and limitations.
• Recruiter brief — Concise role alignment and interview-ready review path.
• External validation kit — Ethical usability-study protocol.

Limitations: Commercial availability does not imply verified revenue, customers, active users, purchases, or completed acquisitions. Deterministic scores are heuristics rather than certified predictions. Local-first storage is not automatically encrypted, durable, or compliant.

If you work in product, software, technical operations, research, agencies, or local-service web implementation, I would rather hear what is confusing, unnecessary, missing, or commercially weak than receive generic praise. Thoughtful criticism, bug reports, workflow objections, and real-world use-case feedback are welcome.

Follow along here on GitHub, review the Five-minute guided walkthrough, or connect with me on LinkedIn.