[2/2] Nubank. (Category Business)
Continue. story About Nubank, I wanted to summarize the theses of Vitor Oliveira, who talked about technology at Nubank on the podcast.Hippsters Ponto Tech #459" (Podcast in Portuguese) April 2025 year, and in August. 2025 He was replaced by Eric Young. (Ex-VP Engineering Snap, Google, Amazon)This, in my opinion, gave a clear signal to investors that Nubank is ready to scale its engineering branch for the global expansion discussed in a previous post.
1Tests as Nubank's 'first line of defense' Nubank has a big focus on testing: unit tests, integration, multi-service tests and experiments with genAI test generation The rationale is that for big fintech, it’s not a “nice to have” but a “must have” to protect customers’ sensitive financial data and keep the bar high on stability, security and privacy.
2The evolution of architecture: from JVM stack to cloud-first Historically, Nubank was built around a JVM stack. (Clojure + Kafka)which set a certain style of architecture and engineering culture (You can see more about the story in the post.10 years of engineering at Nubank")
- Now the guys have a strategy. cloud-firstBut not "cloud at any cost": they consciously balance between cloud and on-prem, taking into account three criteria: cost, control and essence of the business.
3Key Engineering Principles Vitor emphasized the fundamental principles that guide technical solutions.
- Immutability Immutable infrastructure and services
- Standardization Uniform approach instead of zoo technology
- Continuous Delivery Continuous delivery and secure releases
- Minimizing complexity conscious resistance to the "architectural enthusiasm" that inflates the system
4Company as a system It's not just about technology.
- Nubank looks at itself as if it were holisticCulture and people are as important as code and clusters.
- It's important to scale. teams and processes global development (New markets, different regulations, distributed teams)It's not just about throwing microservices.
- Vitor mentions. Conway’s LawHow Communication Structure Affects Architecture, and Why You Can't Get Good Architecture Without the Right Organizational Design
5Generative AI - a cautious, pragmatic approach Nubank already uses gen AU, but very carefully: they are not chasing hype, but looking for it. Specific, measurable problemswhere AI really improves efficiency They have separate budgets and teams that experiment with AI approaches in testing and other tasks, systematically sifting out what doesn’t take off.
- Main filter: Data security and privacy
6The main idea of the release is something like this. Mature fintech It is not built around fashionable technology, but around.
- Strict engineering discipline (tests, standards, CDs) Conscious architectural solutions (JVM kernel, cloud-first, but with an understanding of limitations) A systematic view of organization and culture Careful, pragmatic implementation of AI, where safety and value are more important than demonstrations of “magic”
P.S. It will be interesting to compare Vitor’s theses with the performances of Eric Young, the new CTO that came out at the end of the summer.
#Software #Engineering #Management #Architecture #Processes #SRE #DevOps #Leadership