Ask Claude about me

I built a Claude-powered
AI product from scratch.

Ask anything about my background, projects, or why I want to work at Anthropic. This widget is the work.

Conversations stored in your browser only — never sent to our servers.
Flagship build
Invisible Architecture
AI-powered manipulation detection platform — built solo from zero
↗ LIVE & REVENUE

Full-stack AI SaaS product built in approximately 11 days core, ~3 months to full ecosystem. No prior coding background. Tools: domain expertise, systems thinking, and Claude as a force multiplier. Everything is verifiable — click any link.

FastAPI Anthropic Claude API Supabase Stripe Railway GitHub Python HTML/CSS/JS Prompt Engineering Prompt Caching
The product's 93% margin is only possible because of aggressive prompt caching strategy. The manipulation detection taxonomy — 70 original tactics — is loaded as a cached system prompt. At scale, this collapses token costs by over 80% on every analysis call. Designed this architecture before writing a single line of backend code. The unit economics were the product specification.
From blank repo to live product with working auth, payment processing, AI analysis engine, and conversation management in 11 days. No team. No technical background. The approach: define the system architecture first, use AI to fill implementation gaps, iterate fast. This is the methodology — not the exception.
Free / Starter / Professional / Enterprise tiers. Pricing designed around the token cost model — usage caps at each tier align directly with Claude API cost structure. The 93% gross margin holds at current pricing because the prompt caching solve reduces COGS to near zero per analysis at volume.
The IP behind the product. A structured taxonomy of 70 manipulation tactics derived from 15 years of observing high-stakes human behavior in clinical and administrative settings. Not scraped, not generic — domain expertise converted into a detection framework. This is the moat.
11
Days to MVP
~3mo
Full ecosystem
93%
Gross margin
70
Tactic taxonomy

Published independently via Amazon KDP. Available in paperback and Kindle. The book is both a standalone product and the intellectual foundation for the SaaS platform — the taxonomy originated here before it was engineered into a detection system.

Writing a book while simultaneously building the software product it's based on is not a normal sequence. It required holding two parallel threads: the conceptual framework for a general audience, and the technical specification for an AI classifier. Both had to be internally consistent.

invisiblearc.ai — Production SaaS. Four-tier pricing. Stripe payments. Supabase auth and database. FastAPI backend. Claude API integration with prompt caching. Deployed on Railway.

emptyjay.net — Parent brand and creative tech studio. The umbrella for all products.

E-commerce — Shopify + Printful merchandise store. In progress.

Google Ads — Running paid acquisition campaigns. Managing creative, targeting, and budget allocation independently.

Legal/Business infrastructure — LLC formed. Business insurance. Copyright registered. Trademark applications filed.

Four years building zero-budget data infrastructure at the Kansas City VA Medical Center. Every project was a constraint-driven engineering problem: no budget, no dedicated IT support, no tolerance for downtime.

Lab surveillance system — Eliminated a six-figure annual vendor contract by engineering an in-house replacement using SQL, Power BI, Power Apps, SharePoint, and Power Automate. Built an interim Palantir EMR solution in 2 weeks when the primary contract terminated abruptly. Full production deployment on deadline.

HAI tracking automation — Replaced manual Excel reporting with automated SharePoint + Power BI dashboards. Daily auto-updates. Unit-specific views. Saved hours of staff time weekly.

ICRA workflow digitization — Transformed a manual email process into a LEAF workflow system. Near 100% compliance. Real-time Power BI dashboard with JSON data integration.

This is where the data systems thinking comes from. Production constraints, real users, zero margin for error.

15 years as a registered nurse across surgical inpatient, home health case management, long-term care leadership, and federal infection control management. The clinical career is not separate from the technical work — it is the source of the domain expertise and systems thinking that makes the AI work credible.

Pattern recognition at scale — Clinical practice is fundamentally applied pattern recognition under uncertainty. Ambiguous data, incomplete information, time pressure. This is directly transferable to AI system design and prompt engineering.

Complex workflow design — Managed infection control programs across multiple facilities. Designed protocols, audited compliance, coordinated cross-functional teams. Systems thinking is the job.

Certified in Infection Control (CIC) — Board-certified 2023–2028. Yellow Belt (Lean process improvement).