Senior AI-Native Backend Engineer

About Livio

What You Will Do

• Design and ship backend systems across Next.js + Supabase + Vercel.
• Review code produced by Claude Code before merge — especially security, auth, payment, and integration code.
• Design database schemas, API contracts, RLS policies, integration patterns.
• Integrate third-party APIs: Guesty, Breezeway, Stripe, ERPNext, Twilio, lead providers.
• Set up monitoring, error tracking, operational dashboards.
• Write tests: Playwright E2E for flows, Vitest for unit logic.
• Spot-check AI-generated frontend work and fix what is wrong.
• Push back when a technical decision is wrong — with evidence.

What You Will Not Do

• Decide stack or vendors. Already decided.
• Manage other engineers. None yet, may grow later.
• Talk to investors or customers. Founder handles that.
• Run sprint ceremonies or write Jira tickets. We work async.
• Build entire UIs from scratch. Claude does most UI work.

Required

• Has shipped at least 2 production systems end-to-end.
• Daily TypeScript user. Comfortable with generics, discriminated unions, strict mode.
• Has shipped at least one production app on Next.js App Router.
• Strong PostgreSQL fundamentals — schema design, indexing, transactions, query optimization.
• Familiar with Supabase or willing to ramp in week 1 (RLS, auth, storage).
• Comfortable with third-party API integration including webhooks, signature verification, retry logic.
• Has done code review on a team. Comfortable rejecting code, asking for changes, explaining why.
• Operates well solo. Can scope, ship, and debug without a peer review buddy.
• Reads code carefully — line by line, not skim-and-approve.
• Public GitHub with recent activity, OR ready to demonstrate work in paid trial week.
• Strong written English. Async communication is our default.
• UAE-based, holding GCC residency, or has prior UAE/Gulf work experience.

Ai Fluency Required

• This is non-negotiable.
• We use Claude Code daily across every repo.
• Active user of Claude, Cursor, or similar AI coding tools. Can show specific examples in the interview.
• Reviews AI output critically. Catches hallucinated functions, wrong API versions, missing error handling.
• Can articulate where AI fails — outdated APIs, niche libraries, multi-file changes — not just where it succeeds.
• Familiar with prompt engineering basics. Knows how context windows, system prompts, and tool use work.
• Verifies AI suggestions against documentation before merging. Especially for security, auth, payment code.
• Comfortable that AI writes the majority of production code. If this raises concerns for you, this role is not the right fit.

Bonus

• Hospitality, real estate, or PropTech experience.
• Stripe integrations. Multi-tenant SaaS. PMS or CRM systems.
• Russian or Arabic language.

How We Work

• Async-first. Slack + GitHub PRs. Few meetings.
• One concern per pull request.
• Plan-before-execute for any task touching multiple files.
• Production deploys via Vercel auto-deploy on main merge only.
• Direct, fast-moving culture. Bullet points over essays.

Compensation And Benefits

• Private health insurance.
• Annual airfare per UAE labour law.
• Standard UAE benefits: gratuity, leave, public holidays.
• Hiring process
• CV + GitHub review. 3 days.
• 30-minute intro call. Three AI fluency questions.
• Final call to review trial output together.
• Offer within 7 days of trial completion.
• Total time from application to offer: 3–4 weeks.

Requirements

• 4+ years backend or full-stack experience
• Shipped at least 2 production systems end-to-end
• Strong TypeScript (strict mode, production usage)
• Strong PostgreSQL (schema design, indexing, performance)
• Experience with Next.js (App Router)
• Experience with Supabase or similar backend-as-a-service
• API integration experience (webhooks, Stripe or similar, retries, auth)
• Has reviewed production code and rejected/approved PRs
• Comfortable working independently (async, no hand-holding)
• Strong experience using AI coding tools (Claude/Cursor) in real projects
• Able to verify AI-generated code against documentation