Full-Stack Developer (Long-Term, Full-Time) — Ship & QA AI-Assisted SaaS Builds

About the role We're a fast-moving startup studio building several SaaS products — a field-service CRM for the trades and software for churches, with more coming. A lot of our code is generated by AI build workflows (Claude Code); your job is to own quality: review and harden that code, fix bugs, make solid architecture decisions, and ship features that run at a premium, scalable level. You'll work on a dedicated cloud dev box (provided) with Claude Code, GitHub, Linear, and Slack already wired up. You won't be starting from a blank page — you'll be taking strong drafts to production-grade. What you'll do Review, refactor, and harden AI-generated code into clean, production-ready features Fix bugs filed by our QA testers in Linear — fast and thoroughly Make pragmatic architecture and tradeoff decisions; flag risks early Build and ship features from our written build plans Keep the codebase scalable, tested, and maintainable Communicate clearly in Slack; keep Linear updated Our stack Required: TypeScript, Next.js (App Router), Supabase (Postgres, Auth, Storage, Realtime), Node, Git/GitHub Bonus (our products also use): Inngest (async jobs), Vercel, Fly.io, BullMQ/Redis, FFmpeg/media pipelines, Mux, AssemblyAI, Cloudflare R2 You're a great fit if you Have strong full-stack experience and can work independently Have taken over an existing/messy codebase and stabilized it (not just greenfield) Are comfortable working alongside AI coding tools (Claude Code, Copilot, Cursor, etc.) Write clear English and communicate proactively Can overlap a few hours with US time zones Nice to have DevOps comfort (Linux, Docker, CI/CD) Experience with multi-tenant SaaS, CRMs, or field-service software To apply — answer these (applications without answers are skipped) Describe a time you inherited a messy or AI-generated codebase and got it stable and shippable. What did you do? What's your experience working with AI coding agents? Where do they help, where do they go wrong, and how do you catch it? **What's your full-time weekly availability and your time-zone overlap with US hours? **Link 1–2 projects you've shipped (repo or live).