AI Systems Integration · Recent work
Five things I built recently
Almost nobody selling AI to a small business has anything running in production they can open on a laptop. These are mine. Every number here was counted from a live system or a primary record, not estimated.
VMAP — vmapdata.ai
A live county map that puts owner, value, zoning, flood zone, schools and elected officials on one parcel click. 10,304 map layers across 966 counties in ten states, every Florida county included; five base layers free, the rest a $49/month subscription billed through Stripe. A weekly scan of 270 counties finds new construction, screens it against written rules and writes it into the map; a second pipeline renders marketing video and posts it to four platforms. A 40-point automated audit runs before any county ships. One person built and runs it.
What I built →IBA Consultants — ibaconsultants.com
Nineteen weeks with a 30-year building-envelope engineering firm: 11,000 projects, eight offices. A system of record for clients, projects, activity and follow-up delivered in nine days — it became the firm's working database. A Monday-morning lead sweep across three markets that screens, deduplicates, researches and verifies each new project before writing it, with no human in the loop. Results: 74 projects sourced (25,902 residential units), 7,187 contacts consolidated, 359 logged emails, 47 personalized certificates, zero records lost.
What I built →Kolter Homes — In-house mapping data
Retained since 2024 to source land for master-planned and active-adult communities. Built and maintain a county-by-county landowner database — 180+ tracts, roughly 145,000 acres, 14 Florida counties — shared with the land team monthly; custom Google Earth intelligence maps for 17 counties in Florida and Texas; expansion studies for Texas, the Carolinas and eight further states. The data work is the product; the deals follow from it.
What I built →Remland Commercial — davidremland.ai
This site — listings, a 36-deal transaction history rebuilt from a 2014 archive, annotated aerials for every active property, the contact flow — produced in-house with AI tooling and no developer. It is the demonstration.
What I built →Build-to-Resell Model — West Palm Beach
A spec-home underwriting model for a homebuilder client: buy a lot or teardown, build new, resell. Every active lot and teardown listing in the city — 101 of them — runs through the same cost stack (land, demolition, site, construction, soft costs, contingency, selling costs) and is priced against closed new-construction sales in its own subdivision rather than a zip-code average. 36 had enough nearby sales to support a resale price; one cleared the margin and land-cost tests, at a 45% projected margin. Refreshed from a new MLS export in an afternoon.
What I built →Want the same thing built for your business?
Four fixed-price steps, starting with a two-week audit. The method is on the AI Systems Integration page; the first conversation costs nothing.




