Saratoga Springs, Utah

Dan Piquet

I build practical software, AI agents, and tools for real businesses.

I build production business software and hardware integrations with AI-assisted development. I own the architecture, the integrations, the review, and the deployment.

Selected work

Status as of November 2026.

01

Varda Fieldbook

A phone checklist for Varda Building Co.'s build standard.

Built and deployed to staging. Not in daily production yet.

The build standard was hard to check in the field. Photos, issues, and the phase release needed one record.

I built a phone checklist for 13 phases and 119 checks. Verify a check with photos, log issues, review and release a phase, and export the job. Invite-only accounts, an audit trail, and about 1,800 tests. I built it alone, directing AI coding agents.

  • Next.js 16
  • React 19
  • TypeScript
  • Prisma
  • Postgres (Neon)
  • Better Auth
  • Cloudflare R2
  • Resend
  • Sentry
  • Vercel
Varda Fieldbook on a desktop screen, signed in as the fictional user Alex Rivera. A demo job at 1428 Juniper Lane for Summit Ridge Homes is open on the checklist, with the phase list and a survey check marked verified. The progress line on screen is demo data.

Demo data: Summit Ridge Homes, 1428 Juniper Ln, Alex Rivera. The progress count is not a result.

  • Phone view of Varda Fieldbook phase 01, Planning and feasibility, for the fictional demo job. Two checks are marked verified by Alex Rivera, and another check is still open.
  • Check details in Varda Fieldbook for the fictional demo. A photo is attached, and the history lists verified, photo added, notes edited, and created from Varda standard v1, each stamped to Alex Rivera.
  • Review and release dialog in Varda Fieldbook over the fictional demo project. The dialog asks the reviewer to confirm the phase before releasing it, and says that releasing does not mark unfinished checks complete.

02

Vernops

Scale-house weighing and seed inventory for a family-run seed operation.

Built and deployed. Not in daily production yet.

A family-run seed operation weighed truckloads into lots and bins with a spreadsheet, Apps Script, and paper.

Milestone 1 is a weigh-station app: PIN sign-in, loads that can split across lots, inventory, traceability, bin reports, and CSV. A local bridge reads the serial scale. Offline loads queue and sync once. I directed the work through nine slices.

  • Next.js
  • TypeScript
  • Postgres (Neon)
  • Prisma
  • Tailwind
  • Vercel
Vernops weigh station on a phone-width screen, using fictional sample data. A sample operator has a pending load open for a sample grower. The main scale shows a demo reading, marked as a demo on the screen, with buttons to copy it in as gross or tare. The weight on screen is sample data, not a real ticket.

Sample data. Names and weights on screen are not results.

  • Vernops scale queue with fictional sample loads. Two sample trucks are waiting, each with a control to weigh and complete the load.
  • Vernops traceability screen with fictional sample loads. Filters for date, customer, lot, bin, and truck sit above sample tickets that show which lots were pulled and which bins were filled.
  • Vernops bin report using fictional sample data. A grid of bins and a list of sample loads show where seed went. The weights on screen are demo data.

03

Rehgn

A white-label training and knowledge platform for dealership teams.

Working demo with seeded organizations. Pre-launch. No production customers.

Dealership onboarding is informal. People learn by shadowing, vendor docs do not match the local workflow, and compliance training is split across portals.

Rehgn, first called TORQUE, is the most ambitious system I have built. Companies author their own training and knowledge base. Partners white-label it, starting with dealership operations. 56 features, including SCORM, xAPI, and an AI course architect. Partner-aware tenancy with row-level security, and features gated by entitlements. May through October 2026: 349 commits, 57 database migrations, and about 3,350 passing tests in the last logged run.

  • Next.js 15
  • React 19
  • TypeScript
  • Prisma
  • Postgres (Neon)
  • Tailwind
  • Vitest
  • Vercel
  • Cloudflare R2

04

Sattra

A design-to-installation workspace for interior designers: boards, client approvals, and procurement in one place.

Invite-only beta.

Boards, approvals, and orders were spread across spreadsheets, email, and paper samples. A change in one place did not show up in the others.

Sattra is one workspace from design to installation. Designers build boards from a product library. Clients approve without an account. Procurement and change orders stay tied to those selections, and trades are notified.

  • React
  • Vite
  • TypeScript
  • Fastify
  • Prisma
  • Postgres
  • Tailwind

05

Precision Site Services

Construction purchasing, from a spreadsheet to a custom multi-tenant app.

Purchase-order app is live and still supported.

Material orders, delivery tickets, cost codes, and accounting confirmation were split across a sheet and paper photos.

The live tool is an AppSheet purchase-order app I still support. Field staff confirm orders and upload ticket photos. Project managers enter the cost and cost code. Accounting confirms. A weekend MVP in Cursor grew, with one other developer, into a multi-tenant app.

  • AppSheet
  • Google Sheets
  • Apps Script
  • Cursor

06

AI agent fleet

Product-owner bots, a Pinterest content pipeline, and a résumé bot.

In use on my own work.

Coding, content, and the résumé do not fit in one chat.

I run a personal fleet on my own work. Product-owner bots drive coding agents. A pipeline prepares pins for my Amazon-affiliate Pinterest business. A résumé bot keeps the master résumé. I review what they produce.

  • Claude
  • Cursor
  • ChatGPT
  • Codex
  • Grok

Design and making

Posters and programs for the American Fork Symphony, CNC wood renders for Kaizen Custom Homes, and laser cut work from the shop.

American Fork Symphony

Concert posters and programs, drawn in Illustrator.

Kaizen Custom Homes

CNC wood wall art. These are renders of the ribbed panels, not photographs of an install.

Laser and maker work

A layered mountain shadow box. The instruction sheet is adapted from a CutCute Paper template and prepared as cut files.

Background

Software is the current work. Before that: agribusiness, hardware in the field, and production operations.

  1. Dan Piquet Consulting

    Owner and software developer

    I build software for construction operations and other businesses: estimating, purchasing, and multi-tenant apps. I also connect hardware to the software, including a truck scale at a seed operation. I use AI-assisted development, and I own the architecture, the integrations, the review, and the deployment.

  2. Renaissance Ag

    Marketing, then partnerships, then product implementation

    In four weeks I designed, built, and tested a hydroponic barley sprouting machine for livestock feed trials at Utah State University. I worked with Utah State and Cornell.

  3. Acuitus Ag

    Co-founder and customer success manager

    We built an Arduino-based data-collection stack for ag vehicles. I assembled, flashed, and deployed hundreds of units to farms, and I built the hardware inventory system.

  4. Double L

    Marketing director, production-support tools, and international business development

    I built a Google Sheets and Apps Script system for production orders, and I developed business with a dealership in Western Russia and Kazakhstan.

Education

  • 2023

    DevMountain — Web development

  • 2011

    Brigham Young University–Idaho — B.S. in Agribusiness Science and Technology

Get in touch

If a business is stuck in a spreadsheet, a one-off tool, or a process that should be software, I would like to hear about it.

dan.piquet@gmail.com