← Missions

Senior Full-Stack Developer (AI-First) — Onboarding & Operations Platform

Budget: $25.0 - $100.0 HOURLY / FULL_TIME ⭐ 5.00 (10) United States

next.js, typescript, react-js, figma

Qualifications préférées

  • Expérience : Expert
## About the project We are building a new platform that guides users through a multi-step onboarding and approval process in a regulated services industry. It replaces an off-the-shelf CRM and a large set of spreadsheets with one system: a guided, staged user portal, live per-item status that an internal team updates and users can watch, a manager view of their people, document upload, a basic support ticket, and a clean mobile-first experience. You would be the senior developer building it, working daily with a technical product manager who owns requirements and runs the sprints, alongside a UX designer. We build AI-first, ship something reviewable every week, and keep the stack standard so the client owns it outright. ## What you will build - The user, manager, and internal portals in Next.js and React. - The data model and workflow for staged onboarding and per-item status, with full history and an audit trail. - Live status updates, document upload and storage, email and text notifications, and a basic ticketing flow. - Auth and role-based access, handling personal data securely. - A one-time data migration from an existing CRM and a large set of spreadsheets into the new model. - CI/CD, tests, and monitoring so releases are safe. ## Tech stack Next.js, React, and TypeScript on Vercel. Supabase (PostgreSQL, auth, storage, realtime). Twilio and SendGrid for notifications. Sentry for monitoring. GitHub with CI/CD. Tailwind and shadcn/ui, with Figma for design handoff. Linear for tracking. We use an AI coding assistant (Claude Code) as the primary build tool. ## How we work - AI-first: you drive the AI assistant to move fast, and you review every change. You own the risky parts (data model, auth, security, migration correctness) rather than trusting the AI blindly. - Daily communication with the PM, weekly reviewable deliverables, and UX validated by a power user each week. - The code lives in the client's own GitHub organization, on standard, open components. No proprietary frameworks. ## Requirements - 5+ years building production web applications, with strong TypeScript, React, and Next.js. - Solid PostgreSQL and data modeling, plus hands-on Supabase or equivalent (auth, storage, realtime, row-level security). - Experience handling personal data and building secure auth and role-based access. - Real data-migration experience: messy sources, normalization, and reconciliation. - Fluent with an AI coding assistant and disciplined about reviewing what it generates. - Mobile-first, responsive UI work with Tailwind and a component library like shadcn/ui. - Clear written communication and comfort with a weekly-delivery cadence. ## Nice to have - Regulated or compliance-heavy domains. - Workflow or CRM-style products, notifications, and error monitoring. - AI features shipped to production. ## Engagement - Full-time, roughly 40 hours a week, remote, for about six months to a launched MVP, then lighter ongoing support. - Daily overlap with US business hours for standups and reviews. - Rate is open. We are hiring for judgment and reliability, not the lowest bid. ## How we screen We run a short paid trial task on a real ticket, plus a live technical conversation and a brief pairing session. Every claim below gets a live follow-up, so answer from your own experience. In your proposal, please do three things: 1. Link a project you personally built, then record an unscripted 2 to 3 minute screen share (Loom or similar) in your own voice, walking through one part of the code you are proud of and one part you would refactor now, and why. 2. Describe a production bug you personally chased down: what the symptom looked like, the wrong lead you followed first, and how you actually found the root cause. Be specific enough that we can ask follow-up questions. 3. Tell us about a time your AI coding assistant produced code that looked correct but was not. How did you catch it, and what did you change in your own workflow afterward? Please do not send an AI-written cover letter. We read closely and we will notice. Show us how you think.
Ouvrir sur Upwork

AI proposal draft

Generate a short cover letter for this job. Edit before sending.

Sign in to generate an AI proposal draft.

Connexion