Senior UI/UX Designer (Design Systems)

We are looking for a Senior UI/UX Designer to become the design function for our Generative AI team. We build AI products for web and mobile, and we build them quickly, with AI assistance in the loop. Today the engineers writing the code are also making the design decisions. That keeps us moving, and it shows: every project reinvents the same screen, and nothing looks like it came from the same company.

This role has two halves. 

The first is the design work itself: the flows, the screens, the states, and the visual and interaction decisions across products that are shipping now. The second is the part most design roles stop short of, and it is why this role exists. You will turn your own design work into a design system, tokens, components, and patterns, and hand it to engineering in a form they build from directly.

A folder of mockups does not scale to several products and a team that ships daily with AI assistance. A system does. You will not be asked to be a front-end engineer, but you are expected to take your design all the way to the artifacts engineering builds from, and to use AI tooling to help you get there. The measure of success is how much good design work happens without you in the room, not how much passes through you.

The role spans four areas, and you will move between them continuously:

• Design user flows, screens, states, and interactions for Generative AI products that are actively shipping to internal and external users.
• Design the whole experience rather than the happy path: empty, loading, error, and edge cases, first run and hundredth.
• Establish the visual language for the team's products and hold the line on it: typography, color, spacing, layout, iconography, and motion.
• Prototype where it helps a decision get made faster than a discussion would.

• Find the patterns across the screens you have designed, and decide what earns its place as a token, a component, or a documented pattern, and what stays a deliberate one-off.
• Build and maintain the system in Figma to a standard other people can build from: variables, component properties, auto layout, and libraries.
• Hand the system to engineering in a form they use directly: exported design tokens, the CSS or theme configuration generated from them, and written usage rules with examples.
• Keep the system current as the products change. Version it, say what changed, and retire what nobody uses.

• Design the patterns AI products need and most design systems do not cover: chat and conversational surfaces, streaming output, visible agent steps and tool calls, citations and sources, and human review and approval.
• Design for uncertainty: latency, partial answers, model error, and the affordances that let someone correct or stop an agent.
• Use AI tooling throughout your own work, in Figma, in Claude Design, and in the handoff layer, to get from a design decision to a usable artifact quickly.

• Package the system so AI-assisted work follows it by default, as repository-level rules and skills that coding agents load, so a generated screen starts inside the system rather than beside it.
• Work alongside the engineers as the system is adopted. Review what shipped against what was designed, and fix the system where the system was at fault.
• Keep the whole thing non-blocking. Nobody should wait on you to build a screen.

We care more about how you think and what you have built than about a checklist. 

• Experience: Eight or more years in UI/UX or product design, with real ownership of the design decisions rather than execution against someone else's direction.
• Craft and Judgment: Strong visual and interaction design. Your work holds up when someone looks closely, and you can explain a decision in terms of what it does for the person using the product rather than in terms of taste.
• Systems Thinking: Experience turning design work into something reusable: tokens, components, patterns, or a documented library that other people built from.
• Web and Mobile: Comfortable designing for both, and clear about where a shared system stops and platform convention takes over.
• Figma: Strong Figma skills, including variables, component properties, auto layout, and libraries.
• Handoff: Comfortable producing what engineering builds from, and willing to work into the token and CSS layer. You do not need to be a front-end engineer, and AI tooling can help you close the gap.
• Working with AI: You use AI tools deliberately in your own work and can say where they helped and where you overrode them. This is how the team builds, so it matters.
• Portfolio: Work you can walk us through and talk about. Confidential or unlaunched work is completely fine, and a walkthrough works as well as a link.

None of these are required, and their absence will not count against a strong candidate:

• Experience designing for AI or agentic products.
• Familiarity with design token pipelines such as Style Dictionary or the W3C design tokens format.
• Front-end ability beyond the token layer, for example building components yourself.
• Accessibility practice: keyboard and focus behavior, and screen reader semantics built into components.
• Right-to-left and bilingual interfaces, particularly Arabic and English.
• Motion and interaction design, dense or data-heavy interfaces, or data visualization.
• Usability testing or research to validate design decisions.

• Own the Design Function: You are not one designer among many. You set the visual and experience direction for the team's products, and you decide what the system is.
• Unusual Leverage: Your design decisions do not stop at one screen. Once they become tokens and components, they shape every product the team ships, including the work generated with AI assistance.
• Work at the Front of AI Product Design: You will design interaction patterns for agentic products that most design systems have no answer for yet, and work with AI tooling as a daily part of the craft rather than an experiment.
• Close to the Build: You will sit with the engineers building your designs, see what shipped, and change the system when the system was wrong.