Senior Full-Stack Engineer, Data Product

What we are building
A data product that maps gambling regulation across more than 200 jurisdictions: which markets are open, who licenses operators, how they are taxed, what the advertising and player-protection rules are, and how all of it changes. Today it is a public, content-heavy site with strong organic traffic. Behind it sits a research database, a CMS for editors, and a monitoring system that reads regulators, gazettes and parliaments every day and turns what it finds into alerts and, increasingly, into corrections on the pages themselves.
Two things are changing. The front end is getting a full redesign with a product designer who has just joined. And the product is moving from a public reference to a paid one: user accounts, reports people can buy and find in their account, and a back office to run it. None of that exists yet. You would build it.
What you would own
The web application. Next.js and TypeScript. Implement the redesign with the designer, keep the site fast and indexable, and later add accounts, authentication, payments and purchasable reports.
The data layer. Several SQLite stores with a strict one-way flow from canonical data to mirrors, validation gates that refuse silent data loss, and scripts that apply researched corrections. You would keep these aligned and extend them as the product grows.
The monitoring pipeline. Python services on a Linux server: crawlers for regulator sites, gazettes and parliaments, a matching layer, model-based vetting, and a Telegram bot the team uses to approve findings. You would make it catch more, post less noise, and land approved changes on the site.
AI in the loop. We use models to build and to vet. You would design the pipelines that call them with fixed schemas, verify their output against the source, and treat everything they return as a claim until checked.
Operations. One server, systemd, our own deploy script with rollback, continuous database replication, Cloudflare in front. Deploys must not break the live site, and the person who owns the code owns that too.
The back office. Once accounts and reports exist, the tooling for the team to manage customers, content and data without touching the database by hand.
What we need from you
Senior experience shipping and running production web applications end to end.
Strong TypeScript and React. Next.js specifically, or the willingness to be fluent in it within weeks.
Comfortable in Python for scripts, crawlers and services. Most of the monitoring code is Python and you will inherit it.
Solid SQL and a habit of validating data rather than trusting it. Experience with SQLite is a plus; the discipline matters more than the engine.
Linux server operations: systemd, deploys, backups, reading logs before guessing.
Experience integrating LLMs into a pipeline with structured outputs, and a clear view of why model output is data, not instruction.
A verification mindset. This product is only worth anything if the facts on it are right. You read the primary source, you check the figure, you say when you are not sure.
Clear written communication. Most of our coordination is asynchronous and in writing.

Nice to have

Technical SEO experience. The site lives on organic search.
Auth, payments and subscription systems built from scratch.
Scraping hostile or JavaScript-heavy sites, and knowing when to stop and look for an API instead.
Interest in regulation, compliance or legal data. You do not need to know gambling law; you need to care whether a statute citation is real.
How we work
Small team, direct communication, decisions written down with the reason attached. Nothing in the current architecture is sacred: it reflects choices made along the way, and if you see a better direction for the code, the data layer or how we run things, we want that conversation. AI writes a lot of our code; your judgement decides what ships.