Sharp Generalist Engineer w/ Product Sense — Claude Code, Firebase/GCP a plus

Description: We're a small, fast-moving startup building an AI platform that turns property photos into detailed insurance-claim inventories (computer-vision detection → product matching → pricing → reports). We have a backlog of meaningful work and not enough hands. We're looking for one sharp generalist to own tasks end-to-end. This is not a "just write the code I describe" role. We need someone who can take a loosely-defined problem, figure out what's actually worth building, design a clean solution, and ship it — then tell us what they'd do differently. What you'll do - Pick up self-contained tasks (features, UI/UX improvements, pipeline/data work) and drive them to done - Design simple, user-friendly interfaces and workflows — not over-engineered ones - Make product calls: what's genuinely valuable to the end user vs. what's gold-plating - Work primarily in Claude Code (we provide the subscription / AI tools) - Ask good questions and take real ownership of the code you touch You are - A capable engineer (you don't need to be senior, but you ship working software) - Someone with real product and design sense — you've built interfaces people actually liked using - Part product manager: you can tell a valuable use case from a distraction - A fan of concise, clear, simple design and engineering principles - Proficient with Claude Code (or clearly able to get there fast) - Available with flexibility but also able to commit to regular, steady hours Nice to have (not required) - Firebase / Firestore · Google Gemini models · Google Cloud Platform · business/enterprise pipelines & workflows Logistics - Rate: $10/hour, ongoing part-time with steady work - Stack: Next.js / React / TypeScript, Firebase, Google Cloud, Gemini - AI tooling provided (we prefer Claude) How to apply — Instead of a cover letter, we want to see how you think. We've written up a real problem we've already solved in production and turned it into a short paid test task: 👉 Flow Contents — Engineering Test Task https://docs.google.com/document/d/1ZgOogH40UTdsPpLrZTK-zbY5lQkyZchmcCO_SbZQbXg/edit?usp=sharing The task: given a cropped image of a detected product, find the cheapest and fastest way to retrieve the exact retailer product URL (not a Google Shopping link) — bonus points for a one-API-call solution. The doc has screenshots of our platform, full context, and exactly what to hand in (the API you'd use, the response structure, and the cost for ~1,000 products). To apply: 1. Read the linked doc and complete the task (we estimate 2–4 hours). 2. Reply with a link to your write-up (Google Doc / PDF / Markdown). 3. Start your proposal with the word "crop" so we know you read this. We have an internal answer and grade against it — but we care far more about how you reason than about matching us exactly. Strong submissions move straight to a paid work.