# Building Nubank

We make the extraordinary happen

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Beyond Borders: Why International Talent Chooses to Build a Career at Nubank

DevFeed: [Beyond Borders: Why International Talent Chooses to Build a Career at Nubank](<https://devfeed.tech/articles/beyond-borders-why-international-talent-chooses-to-build-a-career-at-nubank-141922.md>)

Original publisher: [Read original article](<https://building.nu.com/beyond-borders-why-international-talent-chooses-to-build-a-career-at-nubank/>)

Author: Nubank Editorial

Published: 2026-11-18T18:08:14Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Tech Careers](<https://devfeed.tech/topics/tech-careers.md>), [Learning](<https://devfeed.tech/topics/learning.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Developer experience](<https://devfeed.tech/topics/developer-experience.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [accessibility](<https://devfeed.tech/tags/accessibility.md>), [adapting](<https://devfeed.tech/tags/adapting.md>), [after](<https://devfeed.tech/tags/after.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-powered](<https://devfeed.tech/tags/ai-powered.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [alternatives](<https://devfeed.tech/tags/alternatives.md>), [brazil](<https://devfeed.tech/tags/brazil.md>), [build](<https://devfeed.tech/tags/build.md>), [building](<https://devfeed.tech/tags/building.md>), [career](<https://devfeed.tech/tags/career.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [culture](<https://devfeed.tech/tags/culture.md>), [international](<https://devfeed.tech/tags/international.md>), [learning](<https://devfeed.tech/tags/learning.md>), [product](<https://devfeed.tech/tags/product.md>), [product-operations](<https://devfeed.tech/tags/product-operations.md>), [talent](<https://devfeed.tech/tags/talent.md>), [work](<https://devfeed.tech/tags/work.md>), [world](<https://devfeed.tech/tags/world.md>)

### AI overview

Nubank's international teams bring together professionals across markets and functions, with contributors describing career growth, cross-cultural collaboration and learning as key parts of the experience. A colleague who moved from Belgium cited Nubank's entrepreneurship and international expansion; an engineer in Mexico valued compensation, flexible work and technical challenges, and was surprised by the engineering teams' learning culture. The text also notes communication challenges when working across languages, alongside efforts to build shared understanding. Employees describe feedback, onboarding and training, evolving technical challenges, and experimentation with AI tools as supports for development. Careers may grow through broader expertise as well as promotion. Relocation is not required: colleagues can work from their home markets, while language learning, internal communities and everyday cultural exchanges help build connections.

### Source excerpt

From Brazil to the world: how different backgrounds, shared ambitions, and a culture of continuous learning are shaping the way we work at Nubank. The post Beyond Borders: Why International Talent Chooses to Build a Career at Nubank appeared first on Building Nubank.

## Making AI initiatives work for Machine Learning Engineering

DevFeed: [Making AI initiatives work for Machine Learning Engineering](<https://devfeed.tech/articles/making-ai-initiatives-work-for-machine-learning-engineering-132648.md>)

Original publisher: [Read original article](<https://building.nu.com/making-ai-initiatives-work-for-machine-learning-engineering/>)

Author: Nubank Editorial

Published: 2026-11-13T14:39:07Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [machine learning overfitting](<https://devfeed.tech/topics/machine-learning-overfitting.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-research](<https://devfeed.tech/tags/ai-research.md>), [automation](<https://devfeed.tech/tags/automation.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>)

### AI overview

Before committing to an AI initiative, Nubank's PJ Credit Card team asks what workflow or capability it improves, how to measure the gain, and whether parts can be reused. Success measures may include time against a baseline, expanded experiments tied to decisions, or both; quality standards still apply. The example project explores graph-based feature discovery: one phase learns embeddings from existing aggregated features, while another examines transactions and events using a non-graph baseline and a temporal graph. Embeddings are assessed for added value to a risk model, including out-of-time performance, calibration, and stability. Interpretable candidate features can indicate additional signal but do not prove downstream gains. The longer-term aim is to share discovery and evaluation components across risk teams while keeping model-specific inputs configurable.

### Source excerpt

Before building with AI, ask three questions: what problem you're solving, how you'll measure success, and whether the solution can serve others The post Making AI initiatives work for Machine Learning Engineering appeared first on Building Nubank.

## From using AI to operating with AI: What changes in product management when the bottlenecks disappear

DevFeed: [From using AI to operating with AI: What changes in product management when the bottlenecks disappear](<https://devfeed.tech/articles/from-using-ai-to-operating-with-ai-what-changes-in-product-management-when-the-bottlenecks-disappear-106921.md>)

Original publisher: [Read original article](<https://building.nu.com/from-using-ai-to-operating-with-ai-what-changes-in-product-management-when-the-bottlenecks-disappear/>)

Author: Nubank Editorial

Published: 2026-10-29T16:30:40Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Product Management](<https://devfeed.tech/topics/product-management.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [human review](<https://devfeed.tech/topics/human-review.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [product analytics](<https://devfeed.tech/topics/product-analytics.md>), [Pull Request](<https://devfeed.tech/topics/pull-request.md>), [Spec Driven Development](<https://devfeed.tech/topics/spec-driven-development.md>), [Internationalization (i18n)](<https://devfeed.tech/topics/i18n.md>), [Google Calendar](<https://devfeed.tech/topics/google-calendar.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-coding-assistant](<https://devfeed.tech/tags/ai-coding-assistant.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [bottlenecks](<https://devfeed.tech/tags/bottlenecks.md>), [chatgpt](<https://devfeed.tech/tags/chatgpt.md>), [claude-cowork](<https://devfeed.tech/tags/claude-cowork.md>), [product](<https://devfeed.tech/tags/product.md>), [product-management](<https://devfeed.tech/tags/product-management.md>)

### AI overview

Cinthia Scarlatti's examples show how product managers can use AI to reduce execution delays while keeping problem definition, validation and decisions in human hands. For a Mexico-specific customer support correction, she gave Cursor a Jira-style description of the problem, desired behavior and constraints; it proposed options, implemented the selected change and opened a pull request for engineering approval. Reported lead time fell from over a week to under a day. In another case, Cursor used Glean documentation to configure Databricks access, found suitable datasets and analyzed example chats in minutes, helping identify a training issue rather than a technical bug. Scarlatti stresses checking data sources, filters and results before acting. Her Claude Cowork Daily Digest gathers project questions, Slack threads, notes and meeting transcripts into daily summaries, updated documentation and a draft status post; she estimates it saves about an hour per day. The approach depends on clear requirements, relevant company context and review, and she presents the examples as adaptable rather than a universal recipe.

### Source excerpt

How a Lead Product Manager at Nubank turned problem definition into her main tool for working with AI agents, and why judgment stays with the PM The post From using AI to operating with AI: What changes in product management when the bottlenecks disappear appeared first on Building Nubank.

## Beyond the dashboard: What I learned about data without an analytical background

DevFeed: [Beyond the dashboard: What I learned about data without an analytical background](<https://devfeed.tech/articles/beyond-the-dashboard-what-i-learned-about-data-without-an-analytical-background-104683.md>)

Original publisher: [Read original article](<https://building.nu.com/beyond-the-dashboard-what-i-learned-about-data-without-an-analytical-background/>)

Author: Nubank Editorial

Published: 2026-10-28T14:21:45Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [data analyst](<https://devfeed.tech/topics/data-analyst.md>), [Data visualization](<https://devfeed.tech/topics/data-visualization.md>), [dashboards](<https://devfeed.tech/topics/dashboards.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-tools](<https://devfeed.tech/tags/ai-tools.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [analyst](<https://devfeed.tech/tags/analyst.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [background](<https://devfeed.tech/tags/background.md>), [business](<https://devfeed.tech/tags/business.md>), [business-analyst](<https://devfeed.tech/tags/business-analyst.md>), [business-context](<https://devfeed.tech/tags/business-context.md>), [challenge](<https://devfeed.tech/tags/challenge.md>), [charts](<https://devfeed.tech/tags/charts.md>), [code](<https://devfeed.tech/tags/code.md>), [customer](<https://devfeed.tech/tags/customer.md>), [dashboard](<https://devfeed.tech/tags/dashboard.md>), [data](<https://devfeed.tech/tags/data.md>), [data-analytics](<https://devfeed.tech/tags/data-analytics.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [manufacturing](<https://devfeed.tech/tags/manufacturing.md>), [moving](<https://devfeed.tech/tags/moving.md>), [scala](<https://devfeed.tech/tags/scala.md>), [sql](<https://devfeed.tech/tags/sql.md>)

### AI overview

The author's move from manufacturing engineering to senior business analysis showed that learning SQL and Scala took practice but only a few weeks with training and hands-on exercises; the harder skill was explaining "so what?" A dashboard or correct metric is not a conclusion: analysis should interpret what happened and why, then recommend an action and the expected change. In product work, begin with the decision to support, use evidence such as stability metrics or customer experiments, and connect findings to a clear recommendation. The article presents this as the work that turns data into better decisions.

### Source excerpt

How moving from manufacturing engineering into a senior business analyst role taught me that SQL and Scala can be learned quickly, and that the real work starts with "so what?" The post Beyond the dashboard: What I learned about data without an analytical background appeared first on Building Nubank.

## From support to root cause: what a CX person really does at Nubank

DevFeed: [From support to root cause: what a CX person really does at Nubank](<https://devfeed.tech/articles/from-support-to-root-cause-what-a-cx-person-really-does-at-nubank-92854.md>)

Original publisher: [Read original article](<https://building.nu.com/from-support-to-root-cause-what-a-cx-person-really-does-at-nubank/>)

Author: Nubank Editorial

Published: 2026-10-21T12:57:09Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Support](<https://devfeed.tech/topics/support.md>), [AI for customer service automation](<https://devfeed.tech/topics/ai-for-customer-service-automation.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Data analysis](<https://devfeed.tech/topics/data-analysis.md>), [Tool](<https://devfeed.tech/topics/tool.md>)

Tags: [adopting](<https://devfeed.tech/tags/adopting.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-powered](<https://devfeed.tech/tags/ai-powered.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [cause](<https://devfeed.tech/tags/cause.md>), [customer](<https://devfeed.tech/tags/customer.md>), [customer-experience](<https://devfeed.tech/tags/customer-experience.md>), [cx](<https://devfeed.tech/tags/cx.md>), [data](<https://devfeed.tech/tags/data.md>), [first](<https://devfeed.tech/tags/first.md>), [help](<https://devfeed.tech/tags/help.md>), [intelligence](<https://devfeed.tech/tags/intelligence.md>), [job](<https://devfeed.tech/tags/job.md>), [other](<https://devfeed.tech/tags/other.md>), [root](<https://devfeed.tech/tags/root.md>), [solve](<https://devfeed.tech/tags/solve.md>), [support](<https://devfeed.tech/tags/support.md>)

### AI overview

At Nubank, CX work extends beyond resolving each customer case: teams look for recurring causes and either take structured, data-backed findings to product or engineering, or build interim fixes themselves while a permanent change awaits its roadmap turn. In second-level support, an in-house AI-assisted solution built with Google Apps Script sped up manual quality reviews and added reviewer-time metrics. Those measurements helped identify an opportunity to review more cases per agent; the article says a larger sample and correct processes can improve customer satisfaction. It argues CX professionals need analytical skills to find worthwhile problems and the ability to apply AI to test solutions faster. The supplied text is an article excerpt.

### Source excerpt

Why solving the customer's case is only half the job, and how data and artificial intelligence help solve the other half. The post From support to root cause: what a CX person really does at Nubank appeared first on Building Nubank.

## The code was the easy part: Lessons from contributing to the Clojure language

DevFeed: [The code was the easy part: Lessons from contributing to the Clojure language](<https://devfeed.tech/articles/the-code-was-the-easy-part-lessons-from-contributing-to-the-clojure-language-90251.md>)

Original publisher: [Read original article](<https://building.nu.com/the-code-was-the-easy-part-lessons-from-contributing-to-the-clojure-language/>)

Author: Nubank Editorial

Published: 2026-10-19T13:51:12Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Clojure](<https://devfeed.tech/topics/clojure.md>), [experience report](<https://devfeed.tech/topics/experience-report.md>), [Pull Request](<https://devfeed.tech/topics/pull-request.md>)

Tags: [clojure](<https://devfeed.tech/tags/clojure.md>), [code](<https://devfeed.tech/tags/code.md>), [contributing](<https://devfeed.tech/tags/contributing.md>), [debugging](<https://devfeed.tech/tags/debugging.md>), [engineer](<https://devfeed.tech/tags/engineer.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [software-engineering](<https://devfeed.tech/tags/software-engineering.md>)

### AI overview

A senior software engineer recounts contributing to two Clojure core tickets and explains how the work involved understanding user goals, tracing historical behavior, comparing implementation approaches, accounting for compatibility, and clearly presenting the reasoning behind a change. One patch for dynamic library loading in prepl was merged into Clojure 1.13.0-alpha5; a protocol ticket remained under review when the article was written.

### Source excerpt

What two Clojure core contributions taught me about debugging, compatibility, and making technical reasoning understandable to others. The post The code was the easy part: Lessons from contributing to the Clojure language appeared first on Building Nubank.

## Squeezing ~2x MFU out of our training runs

DevFeed: [Squeezing ~2x MFU out of our training runs](<https://devfeed.tech/articles/squeezing-2-mfu-out-of-our-training-runs-88031.md>)

Original publisher: [Read original article](<https://building.nu.com/squeezing-2x-mfu-out-of-our-training-runs/>)

Author: Nubank Editorial

Published: 2026-10-15T17:53:22Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [LLM optimization](<https://devfeed.tech/topics/llm-optimization.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>)

Tags: [ai-research](<https://devfeed.tech/tags/ai-research.md>), [batching](<https://devfeed.tech/tags/batching.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [compilation](<https://devfeed.tech/tags/compilation.md>), [cudnn](<https://devfeed.tech/tags/cudnn.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [faster](<https://devfeed.tech/tags/faster.md>), [model-training](<https://devfeed.tech/tags/model-training.md>), [speed](<https://devfeed.tech/tags/speed.md>), [transformer](<https://devfeed.tech/tags/transformer.md>)

### AI overview

Nubank describes improving training efficiency for a hybrid Transformer model on B200 GPUs through attention-kernel selection, reusable torch.compile output, compiler-friendly Gated DeltaNet integration, and batched Muon updates. In fixed-workload comparisons, model FLOPs utilization rose from 21.1% to 39.3% with AdamW and from 20.2% to 39.2% with Muon, without changing the model architecture or training batch size.

### Source excerpt

How attention-kernel selection, compiler integration, and optimizer implementation shape Transformer model training speed The post Squeezing ~2x MFU out of our training runs appeared first on Building Nubank.

## Feature stores in real-time ML, part 3: Examples from Nubank

DevFeed: [Feature stores in real-time ML, part 3: Examples from Nubank](<https://devfeed.tech/articles/feature-stores-in-real-time-ml-part-3-examples-from-nubank-87076.md>)

Original publisher: [Read original article](<https://building.nu.com/feature-stores-in-real-time-ml-part-3-examples-from-nubank/>)

Author: Nubank Editorial

Published: 2026-10-14T14:36:17Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Training AI Models](<https://devfeed.tech/topics/training-ai-models.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Redpanda](<https://devfeed.tech/topics/redpanda-data-streaming.md>), [clickhouse](<https://devfeed.tech/topics/clickhouse.md>)

Tags: [apache-flink](<https://devfeed.tech/tags/apache-flink.md>), [backend](<https://devfeed.tech/tags/backend.md>), [batch](<https://devfeed.tech/tags/batch.md>), [batch-processing](<https://devfeed.tech/tags/batch-processing.md>), [cdc](<https://devfeed.tech/tags/cdc.md>), [data](<https://devfeed.tech/tags/data.md>), [data-lake](<https://devfeed.tech/tags/data-lake.md>), [data-leakage](<https://devfeed.tech/tags/data-leakage.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [database](<https://devfeed.tech/tags/database.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [dynamodb](<https://devfeed.tech/tags/dynamodb.md>), [real-time](<https://devfeed.tech/tags/real-time.md>)

### AI overview

The article describes three feature-store architectures used at Nubank for real-time machine learning: batch-computed features served from a key-value store, on-demand features queried through Flink and Pinot, and a two-stage Flink pipeline that materializes features in DynamoDB. It explains their trade-offs in latency, freshness, cost, and training-serving skew, and mentions experiments with Feldera and sequential features.

### Source excerpt

Written by: Felipe Almeida, with contributions from Luiz Felix, Vívian Miranda, Lucas Nolasco, and Caroline Custodio. This is part 3 of a 3-part series on Feature Stores for Real-time ML. View part 1 and part 2. Quick recap and introduction In part 1 of this series we described what feature stores are and why they [...] The post Feature stores in real-time ML, part 3: Examples from Nubank appeared first on Building Nubank.

## Beyond A/B Testing: Designing Experiments When the Space Is Too Large to Explore

DevFeed: [Beyond A/B Testing: Designing Experiments When the Space Is Too Large to Explore](<https://devfeed.tech/articles/beyond-a-b-testing-designing-experiments-when-the-space-is-too-large-to-explore-61298.md>)

Original publisher: [Read original article](<https://building.nu.com/beyond-a-b-testing-designing-experiments-when-the-space-is-too-large-to-explore/>)

Author: Nubank Editorial

Published: 2026-09-28T14:21:31Z

Content type: tutorial

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [experiments](<https://devfeed.tech/topics/experiments.md>), [A/B Testing](<https://devfeed.tech/topics/a-b-testing.md>)

Tags: [a-b-testing](<https://devfeed.tech/tags/a-b-testing.md>), [data-analytics](<https://devfeed.tech/tags/data-analytics.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [statistics](<https://devfeed.tech/tags/statistics.md>)

### AI overview

This practical guide explains how optimal design theory can reduce experiment runs when a decision space has many combinations or constraints. It introduces fractional factorial designs and compares them with more general optimal designs for irregular settings.

### Source excerpt

A practical guide to D-optimal designs, Fedorov's algorithm, and how business restrictions shape what we can test and estimate The post Beyond A/B Testing: Designing Experiments When the Space Is Too Large to Explore appeared first on Building Nubank.

## When the team plays, Pix pauses: what the World Cup teaches us about outliers

DevFeed: [When the team plays, Pix pauses: what the World Cup teaches us about outliers](<https://devfeed.tech/articles/when-the-team-plays-pix-pauses-what-the-world-cup-teaches-us-about-outliers-41439.md>)

Original publisher: [Read original article](<https://building.nu.com/when-the-team-plays-pix-pauses-what-the-world-cup-teaches-us-about-outliers/>)

Author: Nubank Editorial

Published: 2026-09-17T15:17:19Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [data](<https://devfeed.tech/topics/data.md>), [context](<https://devfeed.tech/topics/context.md>), [Transactions](<https://devfeed.tech/topics/transactions.md>), [Risk](<https://devfeed.tech/topics/risk.md>), [reliability-engineering](<https://devfeed.tech/topics/reliability-engineering.md>)

Tags: [ai-research](<https://devfeed.tech/tags/ai-research.md>), [context](<https://devfeed.tech/tags/context.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [patterns](<https://devfeed.tech/tags/patterns.md>), [payment](<https://devfeed.tech/tags/payment.md>), [transactions](<https://devfeed.tech/tags/transactions.md>)

### AI overview

The article uses World Cup matches as an example of how coordinated changes in everyday behavior can create outliers in payment data. It explains that instant-transfer volume drops during a match, returns after the final whistle, and may show a halftime pattern, emphasizing that anomalies must be assessed against the normal rhythm of transactions and their context.

### Source excerpt

How World Cup matches turn everyday payment patterns into outliers and what those anomalies can teach us about data, risk, and reliability The post When the team plays, Pix pauses: what the World Cup teaches us about outliers appeared first on Building Nubank.

## Designing Claude for ears, not eyes: An accessibility skill for blind and low-vision users

DevFeed: [Designing Claude for ears, not eyes: An accessibility skill for blind and low-vision users](<https://devfeed.tech/articles/designing-claude-for-ears-not-eyes-an-accessibility-skill-for-blind-and-low-vision-users-41432.md>)

Original publisher: [Read original article](<https://building.nu.com/designing-claude-for-ears-not-eyes-an-accessibility-skill-for-blind-and-low-vision-users/>)

Author: Nubank Editorial

Published: 2026-09-16T14:09:16Z

Content type: opinion

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Web Accessibility (a11y)](<https://devfeed.tech/topics/web-accessibility.md>), [Claude](<https://devfeed.tech/topics/claude.md>), [Users](<https://devfeed.tech/topics/users.md>), [screen](<https://devfeed.tech/topics/screen.md>)

Tags: [accessibility](<https://devfeed.tech/tags/accessibility.md>), [claude](<https://devfeed.tech/tags/claude.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [culture-values](<https://devfeed.tech/tags/culture-values.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [development-environment](<https://devfeed.tech/tags/development-environment.md>), [life-at-nu](<https://devfeed.tech/tags/life-at-nu.md>), [people-with-disabilities](<https://devfeed.tech/tags/people-with-disabilities.md>), [vision](<https://devfeed.tech/tags/vision.md>)

### AI overview

Nubank describes BLIND-CLAUDE.md, an instruction set that adapts Claude responses for developers who primarily use screen readers or read-aloud tools. The approach addresses conventions such as Markdown tables, diffs, diagrams, long logs, and spatial references by adding behavioral instructions without modifying Claude or introducing a native accessibility feature.

### Source excerpt

How a cognitively diverse IT team at Nubank, they coded screen reader-friendly rules into Claude's instructions, treating auditory comprehension as an engineering constraint The post Designing Claude for ears, not eyes: An accessibility skill for blind and low-vision users appeared first on Building Nubank.

## Beyond the data: what a Business Analyst does at Nubank

DevFeed: [Beyond the data: what a Business Analyst does at Nubank](<https://devfeed.tech/articles/beyond-the-data-what-a-business-analyst-does-at-nubank-41431.md>)

Original publisher: [Read original article](<https://building.nu.com/beyond-the-data-what-a-business-analyst-does-at-nubank/>)

Author: Nubank Editorial

Published: 2026-09-15T12:14:41Z

Content type: opinion

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Data analysis](<https://devfeed.tech/topics/data-analysis.md>), [decision-making](<https://devfeed.tech/topics/decision-making.md>), [dashboards](<https://devfeed.tech/topics/dashboards.md>), [Tech Lead](<https://devfeed.tech/topics/tech-lead.md>), [User Experience](<https://devfeed.tech/topics/user-experience.md>), [Finance](<https://devfeed.tech/topics/finance.md>)

Tags: [analysts](<https://devfeed.tech/tags/analysts.md>), [business-analyst](<https://devfeed.tech/tags/business-analyst.md>), [dashboards](<https://devfeed.tech/tags/dashboards.md>), [data](<https://devfeed.tech/tags/data.md>), [data-analysis](<https://devfeed.tech/tags/data-analysis.md>), [data-analytics](<https://devfeed.tech/tags/data-analytics.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [decision-making](<https://devfeed.tech/tags/decision-making.md>), [experimentation](<https://devfeed.tech/tags/experimentation.md>), [meetings](<https://devfeed.tech/tags/meetings.md>), [tech-lead](<https://devfeed.tech/tags/tech-lead.md>), [user-experience](<https://devfeed.tech/tags/user-experience.md>)

### AI overview

This article explains how Business Analysts at Nubank connect data analysis, business context, and experimentation to product decisions. It describes their work in multidisciplinary squads, including investigating metrics, evaluating trade-offs, and making decisions under uncertainty.

### Source excerpt

Understand how Business Analysts at Nubank use data, context, and experimentation to guide product decisions. The post Beyond the data: what a Business Analyst does at Nubank appeared first on Building Nubank.

## From AI demos to real work: How Engineering and Operations learn side by side

DevFeed: [From AI demos to real work: How Engineering and Operations learn side by side](<https://devfeed.tech/articles/from-ai-demos-to-real-work-how-engineering-and-operations-learn-side-by-side-41433.md>)

Original publisher: [Read original article](<https://building.nu.com/from-ai-demos-to-real-work-how-engineering-and-operations-learn-side-by-side/>)

Author: Nubank Editorial

Published: 2026-09-14T16:28:23Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [AI Development](<https://devfeed.tech/topics/ai-development.md>), [software-development](<https://devfeed.tech/topics/software-development.md>), [observability ai agents](<https://devfeed.tech/topics/observability-ai-agents.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-research](<https://devfeed.tech/tags/ai-research.md>), [applied-ai](<https://devfeed.tech/tags/applied-ai.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [development](<https://devfeed.tech/tags/development.md>), [enablement](<https://devfeed.tech/tags/enablement.md>), [technical](<https://devfeed.tech/tags/technical.md>)

### AI overview

Nubank's Ops AI Acceleration Program uses workshops to connect technical AI knowledge with operational expertise. A workshop with PJ Operations introduced generative AI, agents, and the building blocks of AI systems before applying that shared understanding to a real operational challenge.

### Source excerpt

Building shared technical understanding so operational expertise can turn AI into practical improvements The post From AI demos to real work: How Engineering and Operations learn side by side appeared first on Building Nubank.

## Clojure/Conj 2026 Will Offer Free Online Streaming to Expand Access

DevFeed: [Clojure/Conj 2026 Will Offer Free Online Streaming to Expand Access](<https://devfeed.tech/articles/why-this-year-s-clojure-conj-matters-beyond-the-conference-41440.md>)

Original publisher: [Read original article](<https://building.nu.com/why-this-years-clojure-conj-matters-beyond-the-conference/>)

Author: Nubank Editorial

Published: 2026-09-11T17:39:44Z

Content type: opinion

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Clojure](<https://devfeed.tech/topics/clojure.md>), [coding-community](<https://devfeed.tech/topics/coding-community.md>), [Functional programming](<https://devfeed.tech/topics/functional-programming.md>)

Tags: [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [clojure](<https://devfeed.tech/tags/clojure.md>), [conference](<https://devfeed.tech/tags/conference.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [developer-community](<https://devfeed.tech/tags/developer-community.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [events](<https://devfeed.tech/tags/events.md>), [free](<https://devfeed.tech/tags/free.md>)

### AI overview

This editorial explains that Clojure/Conj 2026 will be held from September 30 to October 2 in Charlotte, North Carolina, with free online streaming. It argues that online access can broaden participation by reducing travel and accommodation barriers while preserving live interaction with the event.

### Source excerpt

Making technical knowledge more accessible strengthens the entire developer community The post Why this year's Clojure/Conj matters beyond the conference appeared first on Building Nubank.

## NFR Brain: challenging the status quo of risk management with data

DevFeed: [NFR Brain: challenging the status quo of risk management with data](<https://devfeed.tech/articles/nfr-brain-challenging-the-status-quo-of-risk-management-with-data-41436.md>)

Original publisher: [Read original article](<https://building.nu.com/nfr-brain-challenging-the-status-quo-of-risk-management-with-data/>)

Author: Nubank Editorial

Published: 2026-09-04T14:18:41Z

Content type: opinion

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [risk-management](<https://devfeed.tech/topics/risk-management.md>), [data](<https://devfeed.tech/topics/data.md>), [Risk](<https://devfeed.tech/topics/risk.md>), [Resilience](<https://devfeed.tech/topics/resilience.md>)

Tags: [critical](<https://devfeed.tech/tags/critical.md>), [data](<https://devfeed.tech/tags/data.md>), [data-analytics](<https://devfeed.tech/tags/data-analytics.md>), [data-modelling](<https://devfeed.tech/tags/data-modelling.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [governance](<https://devfeed.tech/tags/governance.md>), [models](<https://devfeed.tech/tags/models.md>), [risk-management](<https://devfeed.tech/tags/risk-management.md>)

### AI overview

Nubank describes NFR Brain, a platform for combining data, modelling, and expert judgement to make non-financial risk signals comparable and actionable. The article presents it as a way to support more consistent, transparent risk prioritization in business decisions without automating judgement.

### Source excerpt

Author : Mayara Zenati At Nubank, we believe the most meaningful problems for customers and for the business rarely come with ready-made answers. That is why we challenge the status quo: not to innovate for innovation's sake, but to remove complexity and build solutions that make important decisions simpler, faster and better. This mindset also [...] The post NFR Brain: challenging the status quo of risk management with data appeared first on Building Nubank.

## How we built a benchmarking framework to horizontally accelerate transaction model research

DevFeed: [How we built a benchmarking framework to horizontally accelerate transaction model research](<https://devfeed.tech/articles/how-we-built-a-benchmarking-framework-to-horizontally-accelerate-transaction-model-research-41434.md>)

Original publisher: [Read original article](<https://building.nu.com/how-we-built-a-benchmarking-framework-to-horizontally-accelerate-transaction-model-research/>)

Author: Nubank Editorial

Published: 2026-09-03T13:53:30Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [Scalability](<https://devfeed.tech/topics/scalability.md>), [data](<https://devfeed.tech/topics/data.md>), [experiments](<https://devfeed.tech/topics/experiments.md>)

Tags: [batch](<https://devfeed.tech/tags/batch.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [data-science-machine-learning](<https://devfeed.tech/tags/data-science-machine-learning.md>), [distributed](<https://devfeed.tech/tags/distributed.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [performance](<https://devfeed.tech/tags/performance.md>), [research](<https://devfeed.tech/tags/research.md>), [training](<https://devfeed.tech/tags/training.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

### AI overview

Nubank describes a benchmarking framework for horizontally evaluating transformer-based transaction representation models across more than 20 benchmarks. The automated, reproducible workflow tests data, architecture, and training changes across multiple tasks and trials, helping identify improvements that generalize. It increased the team's experimentation capacity to roughly five times more experiments per month.

### Source excerpt

The framework that transformed weeks of manual experimentation into automated pipelines for horizontal transaction model research The post How we built a benchmarking framework to horizontally accelerate transaction model research appeared first on Building Nubank.

## When AI skills become supply-chain dependencies

DevFeed: [When AI skills become supply-chain dependencies](<https://devfeed.tech/articles/when-ai-skills-become-supply-chain-dependencies-41438.md>)

Original publisher: [Read original article](<https://building.nu.com/when-ai-skills-become-supply-chain-dependencies-2/>)

Author: Nubank Editorial

Published: 2026-09-02T16:47:57Z

Content type: article

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [supply-chain-security](<https://devfeed.tech/topics/supply-chain-security.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [software-development](<https://devfeed.tech/topics/software-development.md>), [Application Security](<https://devfeed.tech/topics/application-security.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding](<https://devfeed.tech/tags/ai-coding.md>), [ai-research](<https://devfeed.tech/tags/ai-research.md>), [developer](<https://devfeed.tech/tags/developer.md>), [security](<https://devfeed.tech/tags/security.md>), [supply-chain](<https://devfeed.tech/tags/supply-chain.md>)

### AI overview

Nubank describes how AI skills, plugins, MCP servers, agent rules, tool manifests, prompts, and scripts are expanding the software supply chain beyond traditional code. Its security team reviewed more than 2,000 AI skills before distribution, highlighting the need for security controls that address reusable instructions and capabilities in developer workflows.

### Source excerpt

How Nubank vetted 2,000+ AI skills before distribution, building security into the developer workflow without turning safety into a separate gate The post When AI skills become supply-chain dependencies appeared first on Building Nubank.

## Nubank Brazil 2027 Internship Program: developing talent for an AI-native world

DevFeed: [Nubank Brazil 2027 Internship Program: developing talent for an AI-native world](<https://devfeed.tech/articles/nubank-brazil-2027-internship-program-developing-talent-for-an-ai-native-world-41437.md>)

Original publisher: [Read original article](<https://building.nu.com/nubank-brazil-2027-internship-program-developing-talent-for-an-ai-native-world/>)

Author: Nubank Editorial

Published: 2026-08-21T16:36:36Z

Content type: release

Language: en

Sources: [Building Nubank](<https://devfeed.tech/sources/building-nubank.md>), [Nubank](<https://devfeed.tech/sources/nubank.md>)

Topics: [AI Strategy](<https://devfeed.tech/topics/ai-strategy.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [software-development](<https://devfeed.tech/topics/software-development.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [chatgpt](<https://devfeed.tech/tags/chatgpt.md>), [claude](<https://devfeed.tech/tags/claude.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [critical-thinking](<https://devfeed.tech/tags/critical-thinking.md>), [developing](<https://devfeed.tech/tags/developing.md>), [internship](<https://devfeed.tech/tags/internship.md>), [life-at-nu](<https://devfeed.tech/tags/life-at-nu.md>), [talent](<https://devfeed.tech/tags/talent.md>), [technical](<https://devfeed.tech/tags/technical.md>)

### AI overview

Nubank's Brazil 2027 Internship Program is designed to develop early-career talent for an AI-first work environment. The program emphasizes AI fluency, critical thinking, evaluating AI outputs, recognizing limitations, and applying personal judgment. It offers approximately 50 openings and an 18-month development experience.

### Source excerpt

Meet Nubank's new internship program, created to develop talent that combines technical fundamentals, critical thinking and AI fluency The post Nubank Brazil 2027 Internship Program: developing talent for an AI-native world appeared first on Building Nubank.