The work
You are not maintaining a codebase. You are shipping a new application every engagement — a scheduling tool, an inventory system, a support-ticket dashboard, a billing console — each one designed from an empty repo to a running product. The output is judged as production code: real authentication, a normalized schema with sensible relationships and indexes, Redis used deliberately for sessions, caching, or pub/sub rather than sprinkled on, and a Tailwind interface that handles empty states, loading, and errors rather than just the happy path.
Because the applications become training examples, variety matters as much as quality. Two dashboards built the same way are worth less than a dashboard and a workflow engine that make genuinely different architectural choices. Expect briefs that push you across domains you have not worked in, and expect reviewers to care about why you chose server components over client fetching, or a materialized view over a cached query.
What the screen looks for
- Concrete depth in the Next.js App Router: server vs. client components, route handlers, caching and revalidation semantics, streaming.
- Redis judgment — cache invalidation strategy, TTL reasoning, when Redis is the wrong tool.
- SQL design that survives follow-up: relationships, migrations, N+1 avoidance, query plans.
- Evidence you have shipped and operated systems, not just prototyped them. Five years minimum, and the screen will probe whether those years were substantive.
Logistics
Fully remote and asynchronous, contractor basis. Work is scoped per application with delivery windows rather than fixed hours; most contributors run part-time alongside other work, with some taking near-full-time load during active builds. Observed rates on this listing run $100–180/hr, tied to demonstrated seniority and the complexity of the application scope — as reported, not guaranteed.