Moberg has built and run Netgíró's platform since 2013: the backend, the mobile app, and the machine-learning scorecard that cut loan defaults in half. Launch day is the easiest day; the next ten years are the product.
Netgíró is the leading Icelandic payments and buy-now-pay-later platform, offering consumers simple and secure payments online and in-store since 2012. Shoppers pay right away, within a grace period, or in instalments — while merchants are paid in full upfront, with the credit risk sitting on Netgíró's side. That risk is the interesting part.
Moberg has built and operated the platform since 2013 — a year after Netgíró launched. The engagement covers the full product: the backend serving 2,000+ merchants, the consumer mobile app, and, since 2018, the credit-risk models that decide, in real time, who gets a loan. Two focused projects sit inside the long-running partnership: a machine-learning credit scorecard and a Power BI reporting layer for the consumer finance business.
The platform handles user and loan management, multi-channel payments, and automated credit scoring on an event-driven architecture running on AWS. The consumer-facing side is a React Native mobile app; on the merchant side, the platform integrates with major e-commerce platforms including Shopify and WooCommerce, so accepting Netgíró is a plugin, not a project.
The credit-risk scorecard is a machine-learning model built with Azure Machine Learning, Python and R — six months, four FTEs, deployed to production for real-time risk assessment on every credit decision. It was not a one-off: five scorecard generations have shipped since 2018, each feeding audited monthly ECL production and each defended in board meetings. A model you can't explain to a board is a model you don't put in production.
The reporting project — four months, two FTEs — put Power BI dashboards on loan performance, borrower demographics and financial metrics, and automated the financial statements and loan portfolio reporting that used to be assembled by hand.
The scorecard runs inside a live payment flow — a slow or wrong answer is visible at the till.
Every scorecard generation feeds audited ECL production and gets questioned in board meetings. Documentation is not optional.
A platform run since 2013 accumulates integrations, channels and expectations — every change ships into something people use daily.
The payments platform lives on AWS; the scorecard tooling on Azure. The seam between them had to be boring.
The scorecard's headline result is a 50% reduction in loan defaults — the number that decides whether a BNPL business is a lender or a charity. Because the model runs in real time, that risk discipline costs the customer nothing: credit decisions happen at the speed of checkout. Five model generations later, the scorecard is a maintained asset, not a science project.
On the platform side, the value is scale and reliability: Netgíró is accepted at 8 of 10 Icelandic points of sale, used by 10,000+ people daily, with 55,000 active clients aged 21–64 — roughly one in five Icelandic adults. The reporting layer turned month-end from an assembly job into a review job, with automated financial statements and portfolio reporting the finance team can stand behind. Numbers that hold up.
"Netgíró has become an everyday essential for our customers — simple, fast, and always available where they shop. Whether online or in-store, we give people freedom to pay how they want, when they want, with complete confidence."
Runs the event-driven payments platform end to end.
User and loan management, multi-channel payments, credit-scoring services.
The consumer app — one codebase, both stores.
Payment, loan and scoring events decouple the services that consume them.
Training, versioning and deployment of the credit-risk scorecard.
Scorecard development and the statistical work behind each generation.
Loan performance, borrower demographics and automated financial statements.