Software Engineer

The platform is built and going to market now - the agentic roadmap ships this year. The systems you build move real money, contact real people under real compliance constraints, and run autonomously at scale. Youll work directly with the founders: no layers, no ticket factory, genuine ownership end to end. Expanding across NZ, AU, and US.

Full-stack, AI-first, real engineer underneath. You build and ship features across UI, API routes, background jobs, and database. AI coding agents (Claude Code and similar) are a core part of the daily workflow - but you must be a real engineer underneath the tooling. The AI writes a lot of the code; you own the design, the correctness, and the explanation.

• Design, build, and ship features end to end: frontend, backend, schema, tests.
• Use AI coding agents daily to plan, implement, review, and test - and know when not to trust them.
• Review AI-generated and human-written code critically; catch bugs, security issues, and shortcuts before they ship.
• Explain your work clearly: walk the team through what youre building, why it works, the trade-offs you made - demo it on ship.
• Write clear PR descriptions, docs, and async updates others can act on without a meeting.
• Debug production issues across the stack (app, database, background jobs, third-party APIs).

Technical

1-3 years professional software development experience.

Solid TypeScript and React; comfortable with Next.js (App Router) or able to ramp fast.

Working knowledge of SQL and relational databases (Postgres via Supabase).

Understanding of REST APIs, auth, async patterns (queues, webhooks, background jobs).

Testing habits: unit and integration tests without being asked.

Can open the editor and solve the problem yourself when the agent is wrong, stuck, or unavailable. Our interview includes a no-AI coding round.

AI-agent fluency

• Hands-on daily experience with AI coding agents (Claude Code, Cursor, Copilot agent mode) - not just autocomplete.
• Can scope a task for an agent, give it context, review its output line by line, and iterate.
• Can tell when AI output is subtly wrong - understand why - and never merge code you dont.

• Can explain a feature you built to an engineer, a PM, or a customer - at the right level for each.
• Clear written English: PRs, docs, Slack updates.
• Comfortable demoing your work and answering questions about how and why it works.
• Interview includes presenting a feature youve built: what it does, how it works, what youd do differently.

Supabase (Postgres RLS migrations) Tailwind CSS shadcn/ui Zustand TanStack Query Payments / fintech domain Twilio SES VAPI Playwright / E2E testing pnpm workspaces Turborepo

Next.js 16 (App Router) React 19 TypeScript (strict, no any) Tailwind v4 Supabase (Postgres) shadcn/ui Zustand + TanStack Query Inngest Jest + Playwright Railway.

Strict quality gates: no any, typed data access, full build and tests before merge. Async-first, light meeting load. We handle sensitive financial data and real-people communications, so we build with care - and hire people who feel the weight of that.

Location - Remote (NZ / AU / US) .