AI-enabled software engineer for web app build

About the work We're a small, fast-moving team building an AI-native SaaS platform for the e-commerce space. Think: durable agent workflows that take a goal, plan it, execute multi-step work against real third-party systems, and hand a reviewable result back to a human for approval. The domain is high-value and the product is deep - but the day-to-day engineering is what most of this role is about. We move in tight sprints, prioritise ruthlessly, and ship. We're looking for one engineer who can take an idea from a whiteboard sentence to a tested, deployed feature, fast, without dropping our quality bar. What you'll do - Turn roadmap ideas and loosely-specified concepts into shipped, production-grade features - Own features end-to-end: plan → build → test → PR → deploy → verify in prod - Drive sprint velocity with us, and bring your own ideas to the table - we want a thinking partner, not just a pair of hands - Use AI-assisted development (Claude Code, Codex, Gemini CLI,) as a force multiplier, while staying fully accountable forwhat lands - Keep the codebase clean as it grows: refactor, delete, simpl Tech stack - Frontend/App: Next.js (App Router), React, TypeScript (strict), Tailwind - Backend: Node/TypeScript, Postgres + Prisma, durable/queued background jobs for long-running work - AI layer: LLM orchestration, multi-step agent workflows, tool/function calling, structured outputs, sandboxed code execution - Integrations: several third-party APIs (analytics, commerce,agement), OAuth/service-account auth - Infra: Vercel, serverless functions, feature flags, CI (build + lint + typecheck + tests) - Workflow: Git (trunk-based, short-lived branches, conventiond review You don't need every box ticked, but you should be genuinely Next.js/Postgres core and comfortable with LLM/agent plumbing. What we need from you - Velocity with judgement. You ship quickly and you know when to slow down for the thing that matters. - Prompting fluency. You know how to get high-quality output ftasks, giving models the right context, verifying andcorrecting their work rather than trusting it blindly. - Non-negotiable quality habits: git hygiene, meaningful tests (unit + integration), self-reviewed PRs, clean typed code, and QA before you call something done. AI writes a lot of the code; you are accountable for it being correct and secure. - Ownership. You move sprints forward, flag risks early, and d - Security awareness. We handle credentials and make changes to external systems - least-privilege, validated inputs, and human approval on irreversible actions are part of how we build. Nice to have - Experience shipping agentic / LLM-heavy products (not just chatbot wrappers) - Shopify, GA4/Google Ads, or e-commerce data experience - Background job / durable workflow systems (Inngest, Temporal, or similar) - Comfortable in a small team where you wear several hats How to apply Generic proposals will be skipped. To be considered, include: 1. 2–3 examples or case studies of web apps you've built - ideally with AI-assisted development. For each: your specific role, the stack, and whatshipped. Links to live products or repos strongly preferred. 2. How you use AI in your workflow. Which tools, and one concrete example where AI meaningfully sped you up - and how you kept quality high (tests, review, what you did when the model got it wrong). 3. A time you took a vague idea and shipped it fast. What was ambiguous, how you resolved it, how long it took. 4. Your availability (hours/week, timezone overlap) and hourly rate. Optional but a strong signal: a short Loom (2–3 min) walking through one project and how you'd approach working with us. How we'll hire Shortlisted candidates do a short paid trial task (a small, real feature or fix) so we can see how you actually work - velocity, code quality, communication, and how you wield your AI tooling. Strong trials convert to ongoing work quickly.