Madrid or remote. Full-time. Reports to the Head of Applied AI.
You build the rails the agents run on: serving, orchestration, evaluation harnesses, observability and cost control.
What you own
• The infrastructure that runs agents across the portfolio: routing, tool calling, memory, queues, fallbacks.
• Evaluation pipelines and the dashboards that tell us what changed and why.
• Cost and latency per task, and the levers to move both.
• Reliability: agents that fail loudly, roll back cleanly and never lose a user's data.

What you bring
• Five or more years in backend or ML infrastructure; Python plus one systems language.
• Production experience with LLM APIs, vector stores and orchestration frameworks, and the scars to show for it.
• Cloud, containers, CI, and an opinion on observability.
• A bias for boring, well-tested systems.

Why Newry
Own an outcome, not a backlog: a product, a market or a line of the P&L; is yours. Work with agents from day one: they draft, test,



localise and report; you decide. See it land: your work reaches people in 170+ countries, on the first alarm and the last scroll.
How we hire
Five steps, about four weeks. Apply with your work, not a cover letter. Thirty minutes with the hiring lead. A real case from the portfolio, paid if it takes more than a day. Meet the team, and one person you would not work with. Offer written the same week.
Where we work
Madrid, Istanbul, New York, Helsinki, or wherever you do your best work. Remote-friendly across time zones; twelve offices when you want a desk.
About Newry
Newry Integral Media runs apps, services and content used by 100M+ people a month in 170+ countries, for itself and for partners in telco, banking, fintech, airlines, media and retail. 120+ people in twelve offices, agents on every desk. Headquartered in Madrid. newryglobal.com

📌 Applied ML Infrastructure Engineer (Madrid)
🏢 NEWRY GLOBAL MEDIA
📍 Madrid