Senior Full-Stack Developer (TypeScript, Python, Java, SQL)

We're a Dutch e-commerce company. We buy consumer electronics from suppliers and sell them on a large Benelux marketplace. The business runs on a set of in-house tools for purchasing, catalog matching, market data and finance. We're looking for a senior full-stack developer to build four features across four of these codebases, and to keep working with us after that. Each feature has a detailed written specification with the data model, business rules, edge cases, worked examples that double as test cases, and acceptance criteria. THE FOUR FEATURES 1. Purchase-price costing from purchase orders Stack: Next.js, Supabase (PostgreSQL) Automatically fill in the purchase price of sold order lines from the purchase orders in another internal app, which you read through a read-only database connection. Stock is tracked per product and consumed newest purchase order first. Returns go back into stock, and prices are assigned to orders in chronological order. The calculation is a pure, unit-tested function. The writes are atomic and concurrency-safe in Postgres: stock is never consumed twice, an existing price is never overwritten, and reprocessing an import gives exactly the same result. 2. Monthly sales-data import Stack: Java, Spring Boot, MySQL, Playwright, React/TypeScript A scheduled job that fetches monthly sales estimates per product from a third-party SaaS dashboard through its JSON API (JWT auth). It has a Playwright browser fallback, throttling, a status report per product, a configuration screen and a "fetch now" button on the product page. 3. Product matching: categories configurable from the UI Stack: Python, SQLAlchemy, Alembic, MySQL, React/TypeScript Our matching engine (about 12,000 lines of Python) links rows from supplier price lists to our catalog, but it is hard-wired to one product category. You'll turn that into a category registry with generated SQL adapter views (column mapping, unit transforms, safe against injection), a cache and thresholds per category, a new attribute in extraction, scoring and conflict rules, category routing in the LLM-based price-list parser, and an admin page. The existing regression baseline must stay identical. 4. Purchase-proposal engine Stack: TypeScript, Node.js, React, PostgreSQL (Drizzle), MySQL data warehouse, Claude API A deterministic engine that decides per product how many units to buy, from which supplier, for which warehouse and for which of our companies, within several budgets and with deadline and stock limits. It uses greedy fills, a repair step and local search. The proposal is written into an existing planning screen in one transaction and can be undone in one step. An LLM writes a short summary per company, and the code checks every figure in that text against the computed results. Covered by unit tests and seeded property-based tests. This feature depends on work another developer is doing, so it comes last. We'd like to start with feature 1 or 2. Both are self-contained, which makes them a good way to get to know each other. Please navigate to the following link: https://en.wikipedia.org/wiki/Information_technology_management On this page, locate the section titled “IT management disciplines” and click on the first bullet point. On the page that opens, find the section titled “Alignment models” and use the title of the article described there as the first word of your proposal. WHAT YOU BRING • Senior full-stack experience across TypeScript, Python and Java. Expert in at least one, and comfortable working in an existing codebase in the other two. • Strong SQL in both PostgreSQL and MySQL: schema design, migrations, views, transactions, locking and concurrency. • You can implement an algorithm exactly as specified and prove it with tests. • Testing discipline: unit, property-based, integration and E2E tests, with lint, typecheck and tests green in CI. • React and TypeScript frontends that follow an existing design system. • You follow a detailed spec precisely. When a rule is unclear or doesn't hold up against real data, you ask us before filling it in yourself. • Clear written English. NICE TO HAVE • Next.js and Supabase (RLS, Postgres functions, PostgREST) • Spring Boot and Playwright • LLM API integration (structured output, validation, fallbacks) • Docker and Linux servers • Experience with marketplace selling, inventory or purchasing HOW WE WORK • You deliver each feature as a pull request with code, migrations, tests and updated documentation. • We review the PRs and run migrations and deployments ourselves. For server changes you deliver instructions. • Each repository documents its own conventions, and reading them first is part of the job. • Hourly contract with flexible hours. We're looking for a long-term collaboration, and there's more work after these four features. • Communication in English, mostly async, with some overlap with Central European Time.