AI Builder for Groupon
Most companies talk about AI. Groupon is rebuilding how the whole company works, with AI agents at its core. Groupon has said it plainly: "We are rebuilding Groupon as an AI-native company." Not a pilot, not an innovation lab on the side. AI agents in the core of every function, smaller teams, fewer layers, and people who own a problem acting in hours, not weeks. That rebuild needs builders. We are putting together a new team of AI builders who drive it, and we are hiring the first of them now.
What you will actually do
You own one real business area: recruitment, supply, pricing, merchant operations, go-to-market or finance. You work with a sponsor who knows that area inside out. You take a problem nobody has fully defined yet, decide what to build, ship it, and show the number it moved. A typical week: on Monday someone describes a manual process that eats two people's week. By Thursday a first agent-built version runs on real data, and you are showing them what changed.
Two ways in
AI Builder, Engineering. You turn business problems into running, verified software at AI speed. You write the spec and the tests, your agents write the code, you verify and ship. You run several coding agents in parallel (Claude Code, Codex), and every step you repeat becomes an agent or a script. Based in Prague, with trips to Madrid when the work needs it. You report to the tribe lead of the area you join.
AI Builder, Business. You know how a business process really works, and you have already started rebuilding your own work with AI. You map a process, remove the steps that should not exist, and rebuild what is left with agents. You do not need to be an engineer. You do need to show us what you built. Based in Prague, reporting to Radek Hampl, Chief of Staff to the CEO.
Who we are looking for
You build with AI agents every day, at work or on your own projects, and you can show it. You ship. Something working every two weeks beats a perfect plan for next quarter. You learn a new area or stack in days, and you drop a tool the moment the models make it obsolete. You can say what you would not trust in your own work, and you have built things you decided not to ship. No CS degree, set years of experience or specific stack required. Self-taught is fine. Not for you if you wait for a finished brief, enjoy complexity for its own sake, or mostly describe your work as alignment and stakeholder updates rather than what you built and decided.
How we work
One owner per outcome. If what you built breaks, your name is on the fix. Ship fast, learn faster. Days are the delivery unit. Only what moves the metrics that matter. Simplify first, then automate. Every claim you make, you can reproduce.
How to apply
We don't want your CV Send us links to two or three things you have built: a repository with its history, a product, an agent setup, an analysis or a process you rebuilt. For each, one line on what you did, what the team did and what AI did. Add a short walkthrough of one of them, as a 3 to 5 minute screen recording or a few written paragraphs. Nothing confidential from a current or past employer. If it fits, we start with a 15-minute chat. Then a case study on a real Groupon problem (you pick one of two or three, and you can do it over a weekend), one round of feedback, and a session where you show us how you work and defend your decisions. People make every hiring decision.