Data and AI culture: How Nu’s philosophy became a competitive advantage

At Nu DS & MLE Meetup #6, hosted by Nu Mexico, we had the opportunity to hear from Moisés Rojo, a physicist and data scientist with more than 12 years of experience. His journey at Nu includes several milestones: he was responsible for building the first credit models in Mexico, led the creation of the global machine learning infrastructure for credit line increases, and today heads the Global Data Science team for Collections.

Before joining Nu, he served as Director of Data Science Innovation at Kantar and also collaborated with media outlets on data journalism projects.

In his talk, Moisés shared a perspective that goes beyond technical matters and dives into something deeper: the data and artificial intelligence culture that underpins Nubank. More than algorithms and models, it is this philosophy that creates a competitive advantage, enabling the company to simplify the lives of millions of customers in a radically different way.

Technology in the service of customers

Nubank’s central thesis is clear: technology generates efficiency, and that efficiency is returned to customers in the form of better products and fairer prices. But in the field of data science, technology translates into decision systems. Every choice along the customer journey—from the initial analysis to grant a card to defining collection strategies—is supported by machine learning models.

These models act invisibly but decisively. When applying for a card, for example, it is artificial intelligence that estimates credit risk, sets limits, and approves the request in seconds, without bureaucracy or human intervention. The same logic applies to fraud detection, ensuring safety from the very first contact with the customer.

Once the relationship is established, data continues to enhance it. Questions sent through the app are answered quickly thanks to natural language processing. And if speaking with a person is necessary, the system automatically routes the request to the most suitable agent. The same intelligence that approves credit or prevents fraud also personalizes customer support.

As customers mature in their card usage, models continue to work: they analyze spending patterns, calculate default probability, and help decide when and how to raise credit limits. Even collections become less stressful, with strategies tailored to each customer’s reality. The same philosophy applies to other products, such as personal loans. At every stage, technology is there to reduce complexity and give time back to the customer.

The invisible machinery

Making all of this possible requires an invisible machinery supported by multidisciplinary teams and a robust data platform.

Data engineers transform billions of events into actionable information. Machine learning engineers ensure models are scalable and ready for large-scale application. Data scientists build predictions from these datasets, while business analysts translate those predictions into concrete financial impact.

This model, based on multidisciplinary squads that own their solutions end-to-end, eliminates traditional barriers between areas. There is no isolated, distant technology team: everyone works together to turn data into products that truly make a difference in customers’ lives.

Culture as architecture

If technology provides the foundation, culture shapes how we build and operate this ecosystem. Moisés quoted a phrase that sums up this spirit well: “culture is what we do when no one is watching.”

At Nubank, our data architecture choices directly reflect our values. Instead of centralizing everything into a single team, we opted for a Data Mesh model, where each team owns its domain and has the autonomy to build and remodel. This decentralization increases responsibility but also amplifies innovation, as it gives every squad the freedom to adjust and evolve their own data products.

Autonomy, however, does not mean disorder. While we encourage independence, we also build catalogs and governance standards that allow features, data, and best practices to be shared across teams. In this way, every contribution strengthens not just one product but the entire ecosystem, creating a compound effect that multiplies business value.

Another pillar is smart efficiency. Instead of limiting resources, we prefer to give visibility and responsibility over costs. We believe that those closest to the problem are best positioned to decide. This means trusting every NuBanker to pursue efficiency intelligently, balancing savings with impact.

Scale and impact

The combination of robust technology with strong culture generates impact at an impressive scale: more than 100 machine learning models deployed in the past year alone, over 30 petabytes of data processed, 1 billion events ingested monthly, and a network of 200 data scientists working globally. All of this supports more than 122 million customers.

These numbers are not just technical metrics: they represent automated decisions that cut bureaucracy, reduce financial anxiety, and allow people to make better use of their time and money.

Conclusion

As he wrapped up his talk, Moisés emphasized that technology is the lever that allows Nubank to scale its impact. But it is culture—expressed in values, architectural choices, and the way we trust our teams—that turns this technology into a true competitive advantage.

What sets Nubank apart is not only the adoption of machine learning or artificial intelligence, but the way values, people, and platforms combine to create radically different financial products. Products that simplify our customers’ lives and make their relationship with money fairer and more transparent.

In the end, that’s what drives us: using data and artificial intelligence not as ends in themselves, but as tools of transformation for millions of people.