Product Engineer: Replit Expert (Build + QA + GHL Feature Mining + Requirements/UML)
Budget: $6.0 - $8.0
HOURLY / FULL_TIME
⭐ 4.93 (5)
United States
postgresql, python, api-integration, functional-testing, requirements-specification, automation
FullHouse (tryfullhouse.com) is an AI-powered marketing platform in Southwest Florida, USA — social media, reviews, websites, email, SMS, and an AI phone receptionist for real local businesses (clients include Southern Fresh Farms, Trump Vodka, and Compound Capital Partners). Built on Replit, with a deep written spec library. You'll get repo access and ship real features that answer phones and book jobs for real businesses within days.
The role in one sentence — four jobs, one person
You study GoHighLevel to find what's worth stealing, write the complete requirements + UML for our version (adapted to a platform where clients never log in), build it yourself in Replit, and QA everything ruthlessly — yours and the AI agent's work alike — with raw test output as the only accepted proof.
What you'll do
BUILD: implement features end-to-end in Replit (JavaScript/TypeScript, REST APIs, webhooks, database work): vendor integrations, journey logic, admin UI, fixes across the whole platform. You code to written specs and PASTE RAW TEST OUTPUT — never just "done."
QA: feature-by-feature test passes against specs; regression checklists per release; edge cases and compliance behaviors (opt-outs, quiet hours, approval gates); surgical bug reports (repro steps, expected vs actual, severity) — then you fix them.
MINE GOHIGHLEVEL (prerequisite — we test it): you know GHL
hands-on: workflows, snapshots, conversations, calendars, reputation, LC Phone/Email, the AI suite. Monthly teardowns → which features OUR clients would feel → full adaptation spec for our done-for-you, clients-never-log-in model.
SPECIFY: complete requirements docs in our library's format — user story, data model, UI states (loading/empty/error text), edge cases, compliance gates, 6–10 pass/fail acceptance tests — plus UML user journeys (Mermaid preferred: flowcharts, sequence diagrams, state machines) for every major flow before you build it.
Automate: test tooling in Replit — seed scripts, webhook simulators, checklist runners. You respect the platform's hard rules: fail loud, no fallbacks, no hardcoded values, compliance gates never bypassed.
Must have (ALL tested — missing one = don't apply)
Replit expertise: real apps built and deployed there — agent workflows, deployments, secrets, databases. Links required.
Real coding ability: solid JavaScript/TypeScript, REST + webhook craft (signature verification, idempotent handling), comfortable reading any API doc
Hands-on GoHighLevel administration experience
QA discipline + spec-writing craft — you can attach a spec you wrote and a diagram you made
MarTech API familiarity is a strong plus: Twilio, SendGrid, Smartlead, Retell/AI-voice
Excellent written English · US Eastern overlap for one weekly call
Open job