AI Engineer, Design

Visit our Institutional Page

About the role

We're looking for an outstanding AI Engineer with both a passion for design and experience architecting modern AI implementations.

Nubank's design and research function is roughly 260 people. Over the past year they have been building their own tooling: something like 100 AI-assisted solutions and 30+ local knowledge repositories, each built inside a business unit to solve that unit's problem. Most of them work. None of them talk to each other.

The Design Center of Excellence exists to turn that kind of local invention into systems the whole function can build on. It has design ops and research ops. It does not have an engineer — which means the CoE can identify a need, specify a tool, and then has no one to build it.

This is that hire: the first engineer in the CoE, and the person who architects the systems that let design at Nubank scale. You'll work directly with the designers and researchers who use what you build, and the person who scopes a system is the person who builds it.

What you'll work on

NuBrain — the first priority. A global knowledge system for the design and research function, consolidating the 30+ local repositories that already exist. The hard part is not merging content; it's the context architecture — how retrieval is structured, how token consumption stays sane, and how content reaches people and agents at the moment it's useful.

New tools that enable superlative creative marketing and brand expressions at scale. As Nubank scales globally, we are inventing new ways to express our brand in the product and through our marketing. Instead of hiring fleets of marketing designers in every country, we aim to build tools that enable consistent high quality creative expression to be driven centrally.

Canonicalizing what already exists. Roughly 100 AI solutions have been built locally by designers and researchers. Someone needs to understand what they are, decide what becomes standard for the function, and retire the rest before the duplication compounds.

Deciding the shape of the answer. Plugin, MCP server, agent, Claude skill, or a full platform — that call is yours to make and defend. Part of the job is telling the CoE that the tool they asked for should be something else.

What you'll do

- Architect and build the systems design at Nubank runs on — increasingly agentic ones.

- Sit with designers and researchers, understand the need underneath the request, and turn it into something scoped and buildable.

- Stand up or reuse the infrastructure, write the code, and ship it to the function.

- Instrument what you build. Metrics and observability are part of the deliverable, not a follow-up.

- Own the roadmap for CoE tooling: prioritize, sequence, say no, and keep things moving without a PM behind you.

- Be the technical voice in a design organization — make trade-offs legible to people who don't read code.

What we're looking for

An outstanding, deeply experienced AI engineer. This is an engineering hire first. You've spent years building systems other people build on, and recent years working with AI models on things like token consumption, rate limiting, response latency, MCP servers, and other modern AI engineering challenges.

A track record architecting UX systems. Before AI, you designed and built the foundations of user experience at scale — design systems, component platforms, UI frameworks, front-end infrastructure used by many teams.

A love for design that shows up in your work. Since AI, you've used it to build compelling examples and more sophisticated systems than were possible before. We'd like to see them.

Agentic systems that help design scale. Ideally you've built them — agents, tooling, or infrastructure that let designers do more, faster, and better. This is the strongest signal for the role.

A protagonist. You'll serve a 260-person function as a centralized hub. That means navigating stakeholders, leading projects, and pushing work through an organization that doesn't report to you. You drive the work yourself rather than relying on someone else to run the project.

Evidence you've shipped. Send us GitHub, live tools, demos, and systems you can walk us through. We are not asking for a design portfolio.

Backgrounds that tend to fit

- Design systems and UI platform engineers at design-led product companies.

- Engineers at design tooling companies.

- Engineers who have been building AI and agentic tools for designers — in their day job, in open source, or in public.

- Founders and technical leads who were the engineer behind a design-led product.

What this role is not

- Not a designer who codes. The engineering bar is the bar.

- Not a generalist backend or infrastructure role. UX systems experience is core.

- Not customer-facing mobile product work. This is internal systems and tooling for the design function.

- Not a research or modeling role. (The "design modeling" profile — building the models AI consumes — is a different conversation, later.)