AI Builder

Madrid / London / Prague · Europe-based · Full-time · Occasional travel

This is an unusual role. You'll report directly to the Vice President of Talent at Groupon, a NASDAQ-listed company, with no management layer in between — the person sponsoring the work is the person who can greenlight it and put it in front of the company. The VP is a hands-on product partner and builder: they'll define problems with you, prototype and pressure-test what you build, and clear roadblocks at the top of the org. You own the architecture, the code, and the technical calls, with the autonomy that comes with that. You'd join a small, high-intensity team inside the People function, and what you ship reaches leadership directly.

You'll build the data products, decision-support tools, and automation used across hiring, performance, people analytics, people systems and workforce planning — in a company operating in 13 countries. The work lands in front of leadership, not in a backlog.

What you'll do

 Sit inside the People function and learn how it actually works — not the org-chart version, the real one. Then build.

• Pull apart a workflow, process, a piece of talent data, or a workforce question, and come back with strong external references and the solution you'd build.
• Design the data models behind People processes: the recruiting pipeline, performance cycles, headcount and workforce planning, internal mobility, and more.
• Automate multi-step People workflows end to end — routing, extraction, summarization, drafting, and bounded actions across our systems, with human review where it matters. Production systems the team relies on, not prototypes that demo well and die.
• Ship analytical tools that give leaders answers, not another dashboard to interpret.
• Own problems end to end: define the question, model the data, build the system, measure whether it worked.
• Move across recruiting, performance, people analytics, systems, and whatever else the function needs. You don't stay in one lane.
• Travel occasionally to work with People teams on the ground. 

 What makes you right for this

• You've shipped something real. A product, a tool, a system that ran in production or was functionally complete — with users, feedback, and versions. Not a class project with a polished README. This is the first thing we look at, and it carries the most weight.
• You have real AI depth. You've shipped LLM-backed systems that ran in front of real users, and you can talk concretely about the hard parts — evaluation, grounding, failure modes, latency and cost, and where you chose not to use a model at all.
• AI-assisted development is just how you work; you pick up new tools when they make the output better, not because they're new.
• You build things nobody asked you to build. Side projects, self-started tools, scope you expanded past your brief. High agency shows up in what you did when no one assigned it.
• You think in systems, and you simplify. When someone walks you through a process, you're already seeing the entities, relationships, and data flows. You turn a messy workflow into a clean data model fast — and the model is simple, not over-engineered.
•  You finish, and you measure. You quantify outcomes instead of describing activity — "cut review time 60%," not "helped improve the process" — and you don't leave things at 80% or wait for permission.
• You write clean code. Python or TypeScript is your first language or a strong second. You think in SQL rather than translating into it. Your repos show tests and decomposition, not procedural monoliths.

 How we'll assess you

• No mystery here. We score on what you've built (does it run, did it ship, did it matter), your AI depth, the agency in your track record, your code, and how you operate — ownership, speed, impact, and the discipline to finish. Credentials are a tiebreaker, not the gate.

 What we don't care about

• Your GPA. Your degree, or whether you have one. Your major. Your years of experience. Some of the strongest people we've talked
   to have no full-time work history and three things they've shipped.

 Location and work authorization

• You're based in Madrid, London, or Prague and already eligible to work there. If not, but you're you're exceptional and close on this, get in touch anyway. Work authorization alone won't get you rejected.

 Why this role?

• You could go to a big tech company and own a feature inside a feature. You could join a Series A and build from zero for nobody. Here's what this offers that neither does.
• Direct access, for real. You report to the VP of Talent of a public company, who works the problems with you. There's no chain of managers between what you build and the people who act on it, and no quarterly cycle to get something approved. You ship it, the team uses it, and you find out together whether it moved anything.
• Real, messy complexity. People data is some of the hardest data in the company: it spans 13 countries, lives in systems that don't talk to each other, and carries rules that change by jurisdiction. Your systems have to work on that, not on a clean demo dataset.
• A function that's never had this. Most People teams bolt AI on at the edges. Here you'd build it into how the function actually operates, and much of what you ship will be the first version of its kind at Groupon.

 When you join

• The first stretch is immersion: how Groupon's People function actually runs, the AI stack you'll build on, and the problems already on the table. You ramp through real work rather than a classroom. Expect to be contributing in your first week.

 Compensation

• $70,000–$120,000 base, depending on experience and location
• Equity (stock options in a NASDAQ-listed company)
•  Full benefits
•  Occasional international travel, fully covered

 To apply

 Send your resume and one of the following:

• A link to something you built that someone else uses — a deployed app, an open-source tool, an AI agent, a production system
• A 3-minute Loom walking through a system you designed
• A link to a git repository that represents your work