John Bauer.
John Bauer — builder

I build software that businesses actually run on.

Not demos, not slide decks — real products people open every morning to run their day. I'm the CTO at Quo Vadis AI, and on the side I've built and shipped a handful of apps and sites for real companies. This is the stuff that's live.

john@qvadisai.com / CTO, Quo Vadis AI / Remote-friendly
How I work

I'm not a classically trained coder — I came up through construction, IT, and running operations, and I learned to build by building. I work with AI the way a good foreman works with a crew: I know what the finished thing has to be, I direct the work, and I stay on it until it holds up in the real world. I don't ship demos. Everything below is live, and real people run their businesses on it.

What I've shipped

05 projects

ArborIsh

Crew-management software for tree-care companies — jobs, live maps, chat, and billing in one app.

This is the big one. ArborIsh runs a tree-care company's whole day: the owner schedules jobs and assigns crews, the crew sees where they're headed and clocks the work, everyone's on the same live map, and an AI writes a morning briefing that actually reads the weather and the schedule before it talks. It ships two ways from a single codebase — a web app and a native Android app — and it's a real product with real subscriptions behind it.

Under the hood it's React and Supabase with realtime everywhere: crew pins move on the map as people drive, and the whole thing heals itself when a phone drops signal out in the field. Billing runs on Stripe — four plans, a 14-day trial, and a webhook that issues the license key and emails it out on its own. The AI calls go through a server-side proxy so the keys never ship inside the app — a detail I care about because I've set up enterprise security before.

  • Live crew mapRealtime GPS — only during work hours
  • AI commandTalk or type; it creates the job
  • Daily briefingWeather + overdue + open jobs, auto-written
  • Before / after photosStraight onto the job record
  • Team chatReactions, files, emergency contacts
  • ISA cert badgesAI-drawn embroidered patches
  • Stripe subscriptions4 plans; auto license + email
  • One codebaseShips web and native Android
arborish.app
ArborIsh marketing site — 'Run your crew like a well-rooted operation'
ArborIsh app dashboard with AI daily briefing
the app — owner dashboard

stack  React 19 · TypeScript · Supabase (Postgres · realtime · storage · edge functions) · Google Gemini · Google Maps · Stripe · Capacitor (Android)

Strawberry Patched

A local landscaping company, done end to end — the storefront out front, the crew's whole back office behind it.

The public site brings in the work: services, a before/after gallery, a service-area map. The owners edit every word themselves through an admin I wired into GitHub, so they never have to call me to change a phone number. I did the brand too — the strawberry logo, the app icons, a printed door-hanger for direct mail.

Behind a staff login is the real machine — a mobile app the crew runs the day on: a GPS-verified time clock, job scheduling, equipment tracking, and full invoicing. Quotes become invoices, invoices get paid by card through PayPal, PDFs and receipts send themselves, and the owner gets a dashboard with revenue and profit per job. It's live and taking real money.

landscaping.strawberrypatched.com
Strawberry Patched landscaping homepage — 'A yard worth bragging about'
Strawberry Patched crew portal — admin back-office with revenue, billing, crews
the crew portal — admin hub

stack  Astro (public site) · React PWA (crew app) · Express · Prisma / PostgreSQL · PayPal · JWT + roles · GitHub-OAuth CMS · brand & print design

ALC Pro Solutions

A Tampa pressure-washing company — the site that finds the customers and the portal that runs the jobs.

The marketing site does the selling: a clean before/after hero, plus a chatbot and text-message (Twilio) lead capture so nobody slips through while the crew's out working. It's the front door, and it's built to turn a curious visitor into a booked estimate.

Behind it runs the same field-service platform I built for the landscaping company, re-skinned for a different trade — proof it's a real product I can stand up for a new business, not a one-off. This client works in cash and check, so I left card payments switched off. Same engine, wired to how they actually run.

alcprosolutions.com
ALC Pro Solutions pressure-washing homepage — 'Expert power washing solutions'
ALC Pro staff portal dashboard with AI debrief and weather
the portal — same engine, new trade

stack  HTML · CSS · JavaScript · Twilio SMS · lead-capture chatbot · React PWA portal · Express · Prisma / PostgreSQL

Quo Vadis AI

The site for the AI company I'm CTO of — no framework, no paid CMS, and an admin that commits to GitHub.

I built this one lean on purpose: a fast static site with a genuinely clever admin bolted on. Instead of paying for a CMS, the team signs in with GitHub to edit the copy — two small serverless functions handle the login, and every save commits to the repo and redeploys in under a minute.

The content loads live but falls back to text baked right into the page, so it can't break. Little reliability details like that are the part I enjoy.

qvadisai.com
Quo Vadis AI homepage — 'Where are you going, with AI?'

stack  HTML5 · CSS3 · vanilla JavaScript · Vercel serverless functions · GitHub OAuth · YAML content pipeline

In progress building now

SP Visualizer

Snap a photo of a yard, tap a spot, and get an AI redesign you can hand a client as a proposal.

A landscaper needs to show someone what their yard could look like — fast, and without brush tools that assume design skill the client doesn't have. I built the entire tap-to-select experience first, running on fake data, so the flow was proven before any AI got wired in. The real image models drop into one clean interface when it's time.

Faked todayDrops in later
Region detectionSAM 2, server-side on upload
Image edit (tint stand-in)FLUX.1 Kontext / Gemini image edit
Plant library (~12)Full Supabase catalog

stack  React 19 · TypeScript · Capacitor · Tailwind v4 · Konva canvas · pluggable AI provider layer

The long way here

I didn't come up through a CS degree. Before I was 23 I'd already designed custom construction projects in AutoCAD, run the IT infrastructure for a whole campus, set up enterprise security systems, produced live-streamed events, and managed operations across a set of buildings. None of it was the plan — but it's why I can walk into a messy, real-world problem and build the thing that fixes it. The breadth is the edge.

Google AI Professional Certificate AI Fundamentals AI for App Building Manage Security Risks

Also on the side

Thursday

My own AI assistant — phone, desktop, and a browser extension. It routes each task to the cheapest model that can handle it and keeps its memory in a database with semantic search. A private, single-user tool; I mention it as a sample of how I think about architecture, not a link.

Lead Scout

An internal lead-qualifier for Quo Vadis — one React file, no build step, for a day when shipping fast beat setting up tooling. Sometimes the right tool is the small one.