Senior Full-Stack Software Development Team for Platform Next.js 16 + Supabase + Cloudflare)

We're hiring a senior full-stack team to take over and expand an internal operations platform powering a multi-marketplace e-commerce operation (TikTok Shop today, eBay live-auction coming next). Production is live, real money flows through it daily, and the bar is "PR-level engineering, ship multiple times per day, recovery in minutes not hours." This isn't a greenfield project. You're stepping into a working system with: • 30+ Postgres schemas, ~250 migrations • A Next.js 16 (App Router) admin/operator UI • A fleet of Cloudflare Workers fronting Supabase via Hyperdrive • Direct integrations with TikTok Shop API, Microsoft AX (ERP), Keepa, Apify, and Airtable • Bounded queues, pg_cron jobs, webhook ingestion (~25k events/day), and a live inventory-sync engine What we need help with on an ongoing basis: PRODUCT WORK • Operator tooling: bulk-edit grids, override workflows, audit views, mobile layouts • Activity log + incident-review surfaces • Per-item lifecycle dashboards (inbound → setting up → selling → restocking → done) • TikTok Live host tooling (deck, script generation, real-time stock signals) • eBay marketplace integration from scratch (auction + buy-it-now under a single inventory pool) PLATFORM WORK • Push-safety hard rules: single-listing scope, kill-switch gating, retry-once-fail-closed • Saturation prevention: bounded queues, advisory locks, no per-event fan-out • Cloudflare-first execution: keep heavy work off the DB, keep Postgres as system-of-record • Migration hygiene: pre-validate every migration against information_schema, then dry-SELECT it • Observability: structured issue_log, watchdogs, pg_cron health checks, drift detectors RELIABILITY EXPECTATIONS • Production deploys multiple times per day (Vercel + wrangler) • Recovery within minutes when something goes sideways • Live audit of TT endpoints, ERP, pushes, webhooks before every claim of "fixed" • Operator-visible UI for every state change (no silent writes) • Hard rule respect: no auto-bulk pushes, no operator surprise actions STACK • Front: Next.js 16 (App Router, RSC, Server Actions, nuqs URL state), Tailwind, shadcn/ui, AG Grid, TanStack Table, sonner toasts, lucide icons • Back: Supabase Postgres (RLS + SECURITY DEFINER RPCs), Cloudflare Workers (TypeScript), Hyperdrive, Cloudflare Queues, pg_cron, n8n • Integrations: TikTok Shop API (HMAC-signed), Microsoft AX / Dynamics, Keepa, Apify, Airtable, Google Sheets • Deploy: Vercel (Pro, master auto-deploys), Wrangler • Tooling fluency expected: TypeScript strict mode, SQL fluency, browser DevTools, Playwright, real-time DB inspection during incident response CULTURE • Move fast, but read the architecture docs before you touch anything (docs/HOW-WE-BUILD.md, CRONS.md, project-specific AGENTS.md files) • Bias to verification: live-audit your changes against the database after every push • Bias to simplification: do not introduce abstractions until 3 similar lines exist • Bias to operator trust: every action is named, reversible, and visible • Hard rules are hard: push-safety, AX-read-only, no auto-bulk-anything, signed commits WHO THIS IS FOR • A small senior team (2-4 engineers) who have shipped this kind of complexity before • At least one engineer with deep TypeScript + React (Next.js App Router specifically) • At least one engineer with deep Postgres (functions, triggers, RLS, performance) • At least one engineer with Cloudflare Workers + edge experience • UX-aware engineering — the operator-facing UI is half the product • Comfortable working in a system that has already been production-hardened: small surgical changes over rewrites, defensive migrations over assumptions WHO THIS IS NOT FOR • Solo generalists looking for a side gig • Teams that "design before they verify" • Anyone who would consider running git --no-verify or skipping a pre-commit hook • Teams that view "audit before action" as overhead HOW TO APPLY Send: 1. A short writeup (~1 page) of a similar-complexity system you've shipped 2. Links to production work — bonus for live URLs 3. A short example of a recent incident you handled and the structural fix you put in place to prevent recurrence 4. Your team composition + availability (hours/week) Generic proposals will be ignored. We want to talk to teams who have done the work, not teams who want to learn on our dime. Start: Immediate Engagement: Long-term, ongoing Timezone: US-friendly hours required for live incident response