AI Product Designer (Builder)

Deriv; Cyberjaya; On-site; MYR 120,000 - 180,000; Full time

Job summary

Deriv is hiring a product designer who crafts polished UX and builds production frontend experiences with AI coding tools and our design system. You’ll take problems from idea to spec, from spec to interface, and from interface to working PR, all in one role.

You’ll own the end-to-end loop: define the problem, design the experience, write specs with AI, build the frontend, and ship without handoffs. Join a hands-on craft-focused culture at Deriv.

Qualifications

• 4+ years in product design with strong UX thinking and refined UI execution.
• Able to defend every screen decision and notice details (e.g., 2px off).
• Fluent with AI-assisted coding tools and building frontend experiences from specs.

Responsibilities

• Take product problems from research to polished UI and specs.
• Design and build frontend using AI tools and the design system.
• Write precise specs with AI, build components, and open PRs in GitHub.
• Contribute to AI-native design-system documentation and tokens.

Skills

Product design

UX design

Ship-focused mind

System thinking

Tools

Claude Code

Cursor

Codex

Job description

This is an Individual Contributor (IC) role. You’ll lead through craft and what you ship — not people management.

We’re hiring a product designer who can design polished UX and build production frontend experiences with AI coding tools and our design system. You’ll take product problems from idea to spec, from spec to interface, and from interface to working PR.

Why This Matters

The handoff is dead. Traditional workflows move from design to specification, then into an engineering queue, sprint, build, and release. Design intent gets diluted at every stage. This role removes those handoffs: you define the problem, write the specification, design the experience, build the frontend, and ship it.

We’re betting on a different model: designers with deep craft who build directly, using AI as their engineering partner and a design system as their guarantee of quality. The design system makes AI‑built UI consistent; your judgment makes it good.

The Challenge

You're a designer who ships. Not a designer who only hands off Figma files, not a frontend engineer waiting for a spec, not a PM writing documents about products. You hold the whole thing: the human‑centred design thinking, the UI craft, and the build.

Your workflow looks like this: you take a product problem, sharpen the spec with AI, design the experience, then build the frontend yourself — with AI coding tools such as Claude Code, Cursor, Codex, or similar — using our design system's NPM packages, and open the PR. Idea to working product, one person, days not quarters.

And you don’t just consume the design system. You help make it AI‑native: the component APIs, design tokens, examples, agent instructions, MCP‑style integrations, and documentation that help AI tools build on‑brand, accessible UI by default.

Why Deriv

• You'll ship to real users. A global trading platform with millions of users — your work hits production, not a portfolio mockup.
• You'll build with a real system. Our design system ships as NPM packages across client apps, partner apps, and websites — and you’ll help shape how AI agents use it.
• You'll own the full loop. You will take the work from specification and design through frontend build, pull request, and release. No handoff. No queue.

You’ll join a culture that shares openly. Deriv is where we write about what we’re shipping, what breaks, and what we figured out the hard way.

What You’ll Do

Design like a craftsperson

• Take product problems from research to polished UI, with human‑centred design driving every decision.
• Own the UX of AI‑era products: non‑deterministic behaviour, loading and failure states, trust, and clarity about what the system is doing.
• Sweat the details — spacing, motion, hierarchy, accessibility — because AI can generate UI, but it can’t yet tell good from almost‑good. That’s you.

Ship the frontend yourself

• Write product specs with AI that are precise enough to build from.
• Build with AI coding agents against the design system’s NPM packages — composing real components, not generating throwaway code.
• Work in GitHub: branch, commit, open PRs, respond to review, ship.
• Debug when the build breaks, and know when to ask for engineering input without handing off ownership.

Make the design system AI‑native

• Contribute to the instruction layers that teach AI tools how to use our components correctly.
• Write and refine documentation, examples, and MCP‑style integrations that make "AI builds on‑brand UI" true by default.
• Feed what you learn from building back into the system: missing components, unclear APIs, gaps in the docs.

Compress the loop

• Go from idea to testable product in days. Validate with real users, kill what doesn’t work, double down on what does.
• Your value is measured in shipped product, not documents or mockups.

You’re a great fit if you’re

You have serious design craft. You have 4+ years in product design, with work that demonstrates deep UX thinking and refined UI execution. You can defend every decision in a screen — and you notice when something is 2px off.

You build what you design. You're fluent with AI‑assisted coding tools such as Claude Code, Cursor, Codex, or similar. You're comfortable with React/TypeScript at component‑composition level and at home in GitHub. You don’t need to be a senior engineer — you need to ship working frontend experiences.

You can show proof of build. You have a link, repo, Loom, prototype, or shipped product that proves you personally designed and built something, ideally with AI tooling.

You spec with AI fluently. You use LLMs to sharpen thinking, not replace it — turning ambiguity into clear specs, exploring variations fast, and knowing when the model is confidently wrong.

You own outcomes. When the build breaks, you debug it. When AI output drifts off‑system, you help fix the system so it doesn’t happen again. You act as if company success is your personal responsibility.

Bonus Points

You think in systems. You understand design systems from both sides: as a consumer building products with them, and as an architect making them usable by humans and AI agents alike.

You’ve worked with design‑system implementation. Tokens, component APIs, Storybook, package documentation, usage guidelines, or contribution workflows are familiar territory.

You’ve explored AI‑native product or design‑system workflows. Agent instructions, MCP‑style integrations, AI coding agents, or structured documentation for AI tools are a plus — but we don't expect everyone to know our exact vocabulary yet.

Tech stack

• Design: Figma, human‑centred design methods.
• Build: AI coding tools such as Claude Code, Cursor, Codex, or similar; React/TypeScript; GitHub.
• System: Design system NPM packages, design tokens, component documentation, agent instructions, MCP‑style integrations.
• Ops: ClickUp, Slack.

The Honest Reality

This role is demanding because the ownership is real. You own the design, the build, and the outcome. You’ll sometimes ship at 80% polish because validated learning beats perfect craft — and that may hurt, because you care about craft.

You’ll debug AI‑generated code that almost works. You'll write documentation so agents stop making the same mistake. You'll be one of the few people in the room who understands both the design question and the technical answer.

If you want a clean lane — pure designer, pure engineer — this isn’t it. If you want to be the person who designs the product, builds it, and shapes the system that makes it all scale, it might be exactly your kind of role.