Founder Β· Engineer Β· Builds with AI

GED to Infrastructure Empire.

I'm a self-taught founder running two companies and shipping production systems solo β€” 270+ Cloudflare Workers, nine live platforms, four public APIs, and a property-data pipeline spanning 165M+ parcels sourced straight from government records. AI isn't my autocomplete; it's my co-founder. I also take on select outside builds and Claude Code engagements.

270+Workers in Prod
165M+Parcels Indexed
9Live Platforms
justin@proptech:~
β†’cat the-arc.md
# Nov 2024 β€” ChatGPT. Frustrating.
Looked right, broke in prod.
# Jul 2025 β€” launched LHBUSA.
Switched to Claude + Cursor. Clicked.
# 2026 β€” two companies.
270+ workers Β· 165M parcels
8 sites Β· 22+ subdomains Β· 20+ books
β†’./keep-building.sh
πŸ“¦
valuation-enginepublished on npm
⚑
270+ WorkersCloudflare edge
πŸ—ΊοΈ
165M+ ParcelsPropData API

How I Got Here

No CS degree. No bootcamp. A GED, relentless building, and AI as a teacher that never sleeps.

The False Start

Tried learning to code with ChatGPT. Built things that looked right but broke in production. Freelancers wanted $5K for what I could picture in my head. There had to be a better way.

The Pivot

Launched Local Home Buyers USA on a novation model. Switched to Claude + Cursor and the code actually shipped. Started funding the tech with real estate deal flow β€” building both at once.

The Explosion

Two companies in parallel. 270+ workers, public APIs, a sports-betting model, a 165M-parcel data pipeline. I publicly challenged Opendoor to explain their algorithm β€” their pricing team viewed the post and went silent. So I reverse-engineered it and open-sourced my alternative.

"The gatekeepers in property data β€” CoreLogic, ATTOM β€” charge a toll on public records. I decided to build the pipeline everyone else skips."

β€” Why I went all-in on infrastructure

AI as Co-Founder

Claude is the technical co-founder I couldn't afford to hire. Every error is a lesson, every build gets faster. I ship in days what used to quote at months β€” and I've published 20+ books on how I do it.

Synthesis, Not Gatekeeping

Data is public; synthesis isn't. While competitors hide algorithms and resell public records, I build the pipeline, own the model, and ship the interface. That's the durable moat.

Two Companies, One Stack

Real estate funds the tech; the tech compounds into products. Both run on the same edge infrastructure I built and maintain solo.

Real Estate

Local Home Buyers USA

A nationwide acquisitions company built on a novation/partnership model β€” earning the spread instead of tying up cash.

  • β–ΉNovation-first offers β€” end buyer pays retail to the seller, LHBUSA earns the spread.
  • β–ΉCustom CRM + lead engine replacing off-the-shelf SaaS, with PropData enrichment and Claude lead scoring.
  • β–ΉOpen-source valuation engine on npm β€” the transparent answer to iBuyer pricing.
  • β–ΉMigrated off WordPress to a custom Workers/Vercel stack with an autopiloted content hub.
PropTech Β· Sports Β· Data

PropTechUSA.ai

The platform powering the real estate ops β€” plus a growing set of data and sports-intelligence products that stand on their own.

  • β–ΉPropData API β€” 165M+ parcels from government-sourced public records, with delta + pre-foreclosure feeds.
  • β–ΉPropBetEdge β€” an MLB betting-intelligence platform with an owned Statcast/Poisson model.
  • β–ΉPropSports API β€” multi-sport data on RapidAPI, now callable by AI agents via a remote MCP server.
  • β–Ή270+ Workers across 8 sites and 22+ subdomains, all on the Cloudflare edge.

Production Systems

Real infrastructure handling real users, real data, and real money β€” not demos.

PropBetEdge β€” Sports Intelligence Platform ● live

A full MLB betting-intelligence platform: an owned Poisson model built on nightly Statcast (xSLG, barrel%, EV-allowed, park factors) that generates its own prop odds β€” refreshed 6Γ— daily, zero sportsbook dependency. Daily Diamond & Prime picks with a full public track record β€” wins and losses both, a Park Factor engine, PBEcast, umpire grades, live trackers, and a $29/mo paywall after a 3-day trial. NBA, NFL, and NHL are live in beta on the same model. The thesis in one line: data is public, synthesis isn't.

Poisson model6Γ— daily oddsStatcast pipeline$29/mo Β· 3-day trialMLB + NBA/NFL/NHL beta
PropData API$49/mo

165M+ parcels and 16 live sources (Census, HUD, ArcGIS, FHFA, FRED, Zillow ZORI, Redfin, Realtor.com, HMDA) merged into one call β€” built to replace RentCast, CoreLogic, and ATTOM at a flat $49/mo with no per-call fees. Delta and pre-foreclosure feeds for enterprise partners, derived absentee/vacant-land flags at request time.

165M+ parcels16 sourcesRapidAPIgov records only
PropSports API + MCP Servernew

Multi-sport data (MLB/NFL/NBA/NHL) across 37+ endpoints on RapidAPI at $24/mo β€” the only sports API generating its own Poisson odds from owned Statcast, no sportsbook feeds. Now wrapped as a remote MCP server so AI agents in Claude, Cursor, and beyond call it as native tools. Stateless Worker, service-bound to the REST API, key-gated per plan.

Remote MCP37+ endpoints$24/mo Β· RapidAPIowned odds
Addy API + PBE Sports News APIAPI

Two more published APIs. Addy β€” instant address autocomplete backed by the 165M-parcel store, returning AVM, beds/baths/sqft in a single call. PBE Sports News API β€” live sports news scored for prop-betting impact (injuries, lineups, weather) the moment it publishes, as clean JSON. Four public APIs total, all on owned infrastructure.

Address intelligenceNews scoringRapidAPI
Opendoor Reverse-Engineering + valuation-enginenpm

Publicly challenged Opendoor to explain their pricing. They stayed silent β€” so I reverse-engineered it, exposed the real fee spread, and open-sourced a transparent valuation engine as an npm package anyone can use.

npm publishedOpen sourceValuation
270+ Worker Edge Architectureinfra

8 main sites, 22+ subdomains, 270+ Cloudflare Workers, and 9 live platforms β€” orchestrators, cron pipelines, alert workers, intelligence hubs, and API gateways at <50ms across 300+ edge locations. Supabase/PostgreSQL for data, Vercel + Cloudflare for delivery, all maintained solo.

Cloudflare Workers9 platforms<50ms300+ edge
CRM, Dashboards & Content Enginesops

A custom lead-management CRM replacing paid SaaS, a CEO dashboard pulling live APIs into one view, and an autopiloted intelligence hub that researches, writes, and publishes content on a cron β€” the operational backbone behind both companies.

ReactClaude APIAutomation

What I Ship With

Production-tested, battle-hardened, daily drivers.

Cloudflare Workers
edge compute
Supabase / Postgres
data layer
React
frontend
Vercel
deploys
Anthropic Claude API
intelligence
Cursor
IDE
Stripe
payments
MCP
agent tooling

I Document What I Build

The same methods I ship with, written down β€” plus the start of an earned-media trail.

20+Books Published
700+Essays Live
2Whitepapers
18K+LinkedIn Reach
β–Ή

"The AI Blueprint" hit #206 across all of Kindle within 48 hours of launch β€” zero marketing spend.

β–Ή

Featured in The Globe and Mail β€” first earned media on the GED-to-founder, AI-as-CTO story.

β–Ή

"The Third Kind of Mind" β€” a whitepaper on AI cognition, refined across 13 adversarial revision cycles.

β–Ή

The Claude Command series and The Founder's Stack β€” technical AI titles on building production systems with AI.

Let's Build Something

I run my own companies β€” but I take on a handful of outside projects: MVP builds, API/infra work, and Claude Code training for engineering teams that want to ship at the pace I do.

JE

Ask About Justin

AI-powered β€’ Instant response

Hey! πŸ‘‹ I'm an AI assistant that knows Justin's work β€” two companies, 270+ workers, a sports-betting model, and a 165M-parcel data API. Ask me about his projects, his stack, how he learned to code, or whether he's available for work.
Powered by Claude AI