JS Sênior Expert for Claude Code Refactor

About the project We’re a fintech building the next generation of a freight-factoring and banking platform for the logistics/trucking industry. We’re migrating a mature production system (PHP/Laravel + Flutter) to a modern TypeScript monorepo, and we run an AI-first engineering process built around Claude Code. We’re looking for a senior engineer who is genuinely fluent with agentic AI development to help us ship this refactor to production. What you’ll do • Build and harden backend domains in a NestJS + Drizzle (Postgres) monorepo: factoring/receivables (FIDC), banking (BaaS), payments, identity & RBAC. • Bring the new backend to full parity with the legacy system, endpoint by endpoint, verified against the real product. • Write real integration tests against Postgres (no mocks on money paths) and keep CI green. • Work alongside our AI software factory: author specs, review and critically verify AI-generated pull requests, and fix what the agents get wrong. • Contribute to the Next.js 15 admin web app and (later) the Expo React Native driver app. Must-haves • 5+ years shipping production JavaScript/TypeScript; deep TS (generics, discriminated unions, strict mode). • Strong with NestJS (or a similar DI/decorator backend) and SQL/Postgres (Drizzle, Prisma, Kysely, or raw). • Hands-on Claude Code (or equivalent agentic AI coding) experience in a real codebase — you can drive it, review its output skeptically, and catch where it’s subtly wrong. • Comfortable with monorepos (Turborepo/pnpm), CI/CD, and writing tests that actually catch bugs. • Obsessive about correctness with money (we never use floats for currency). • Professional English and strong async communication. Nice to have • Fintech / payments / lending / factoring domain experience. • Next.js App Router + RSC, Zod, React Native/Expo. • Banking-as-a-Service / PIX / Brazilian financial rails. • Overlap with Brazil business hours (BRT/UTC-3). How we work • AI-first: parallel Claude Code agents do much of the building; humans architect, review, and verify. We want engineers who multiply their output with AI but never ship code they don’t understand. • Quality gates: real-DB tests, code review, green CI before merge. • Async, outcome-focused, low-meeting. Engagement • Long-term contract until the refactor ships — [set: ~20–40 hrs/week]. • Rate: [set your range, e.g. $XX–$YY/hr]. Start ASAP. To apply, tell us: 1. A non-trivial feature or bug you shipped with an AI coding agent — what the agent did, what you had to correct, and how you verified it was right. 2. A TS/Node codebase or PR you’re proud of. 3. (Optional) One thing teams get wrong about AI-assisted development.