Product Designer - UI/UX
- Source: LinkedIn
- Xpin AI
- On-site
- Dubai
- Full-Time
- Mid level
Job Description
Design the interfaces for AI products used by governments and large enterprises across the GCC, at a level of craft that stands next to the best software being made anywhere, in two languages that read in opposite directions.
About Xpin AI
Xpin AI builds sovereign AI for governments, enterprises, and institutions across the GCC. Our products are bilingual at their core, spanning Arabic, English, and regional languages, and ship into both on-premise sovereign environments and cloud. Design here is a genuinely hard problem rather than a finishing layer: our users are making consequential decisions with a system whose output is probabilistic, so clarity, trust, and bilingual fluency are not polish, they are the product.
The role
You own the interface design for products across our suite. This is a hands-on individual contributor role. You will not be writing decks about design, you will be drawing the screens, building the components, and sitting with the engineers who ship them.
The bar is high and worth being explicit about. We are looking for the level of craft where typography, spacing, hierarchy, and motion are decided rather than defaulted, where the interface feels obvious to use and considered up close, and where nothing looks like it came out of a template. Most enterprise software in this region does not clear that bar. That is precisely the opening.
What you’ll design
- Product interfaces end to end. Flows, screens, states, and the edge cases that never appear in a portfolio shot: empty, loading, partial, failed, and the long tail of real data.
- Interfaces for probabilistic systems. AI output is not deterministic. You will design how a system streams an answer, shows what it is doing and why, expresses confidence and uncertainty, and lets a user verify or correct it. There is little established pattern language here, so you will be defining it.
- Real-time and voice experiences. Conversational and voice-driven interfaces where timing, feedback, and interruption are part of the design rather than an afterthought.
- Dense, information-heavy surfaces. Dashboards and analytical views where the job is to make complexity legible for people who use the product all day, not to simplify it away.
- The design system, alongside the Design Lead. Components, tokens, and the typographic and spacing foundations, built in Figma so they translate cleanly into code rather than needing reinterpretation.
- Bilingual design at system level. Arabic and English as genuinely equal experiences: mirrored layouts, correct Arabic typography and numerals, type scales that work across two scripts with different rhythms and no shared italic convention. This is one of the hardest and most interesting parts of the job.
- Motion with intent. Transitions and micro-interactions that communicate state and causality, specified precisely enough to be built as designed.
What good looks like here
- Typography is treated as a primary tool, not a setting. You have opinions about type scale, measure, and optical alignment, and you can defend them.
- Spacing and layout follow a system you can articulate, and the system holds up under pressure rather than being abandoned on the hard screens.
- Restraint. You know what to leave out, and your interfaces do not decorate to appear designed.
- Your files are structured well enough that another designer or an engineer can pick them up without asking you what anything means.
- You design the whole thing, including the states nobody screenshots.
Tool stack
- Core design: Figma at depth, including variables and modes, component properties and APIs, auto layout, library architecture, and Dev Mode for handoff.
- Prototyping: Figma prototyping for flows, with ProtoPie or Framer where an interaction needs real logic, timing, or conditional state to be evaluated properly.
- Design systems and tokens: Tokens Studio for token management, feeding a Style Dictionary pipeline maintained with our frontend engineers. Semantic token architecture, theming, and multi-mode systems that support both language directions.
- Motion: Rive or Lottie for production motion, authored so engineers implement what you designed rather than approximating it. After Effects where the pipeline calls for it.
- Typography and localisation: Arabic and Latin type pairing, variable fonts, type scales that hold across two scripts, RTL mirroring, bidirect…
Skills
- built-in
- Figma
- Typography
- mode
- protopie
- framer
- lottie
- React
- next-js
- Tailwind
- storybook
- Github
- maze-design
- usertesting
You missed this one - but you don't have to keep missing out. Try Club members get matched with open opportunities before they're gone, so you're always first in line, never last.
