This is an Individual Contributor (IC) role. You'll lead through creative expertise, systems thinking, and the tools you build — not people management.
Why This Matters
You'll be producing creative for a global trading platform reaching millions of traders across languages, cultures, and channels. The old model from brief to launch takes weeks. Your job is to collapse that into hours, then automate it so it stays collapsed.
That means moving fluidly across illustration, static, motion, and video — and treating every repeated manual task as a bug to be fixed with a better system.
The Challenge
You're not here to push pixels. You're here to build the tools that make pixels push themselves.
Most designers use AI. You build with it. When a workflow is slow, you don't just prompt harder — you use Claude Code, Codex, Cursor, or similar AI coding agents to build the tool, script, or workflow that fixes it.
You're a designer who can make software work for creative production — using AI agents to build the tools, scripts, and workflows that remove bottlenecks.
Why Deriv
- You'll work in production, not theory. Real campaigns, real deadlines, real volume.
- You'll build, not just brief. Your workflows become internal products — pipelines that ingest market intelligence and output production-ready assets.
- You'll multiply the whole Studio. One-off wins become repeatable engines other designers use daily.
- 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 the factory, not just the product
- Build reusable AI workflows for high-volume output: PPC sets, partner kits, social series, localised variants.
- Create prompt libraries and visual recipes that keep AI output on-brand at scale.
- Turn manual production steps into tools other designers can run without you.
- Design repeatable systems with clear inputs, variables, QA checks, and brand controls.
Build like a creative technologist, think like a designer
- Use Claude Code, Codex, Cursor, or similar tools to build scripts, small apps, and automations for creative production.
- Work in GitHub: branch, commit, open PRs, respond to review, and collaborate with engineers where needed.
- Connect AI models to internal pipelines via APIs, including image generation, video generation, TTS, LLMs, and creative QA tools.
- Build scrappy, reliable tools first — then improve them based on real production use.
Execute across the full stack
- Move comfortably between graphic design, illustration, motion graphics, and video editing.
- Turn a static concept into a 15-second motion piece using AI video tools.
- Use the brand system to keep output consistent while pushing what's possible.
- Know when AI output is good enough to ship, when it needs craft, and when it needs to be rejected.
Design for marketing outcomes
- Understand the campaigns your assets serve — audience, channel, funnel stage, and performance goal — and design accordingly.
- Work with marketing on briefs, but own the craft and the systems behind it.
- Build workflows that make campaign iteration faster without lowering creative quality.
Who You Are
You've produced creative at scale. You've felt the pain of manual production that can't keep up — and you refuse to work that way anymore.
You're a true full-stack creative. No hiding behind "I only do static" or "I only do video." You see visual language as a whole and use whatever medium tells the story.
You build with AI, not just prompt it. You have hands-on experience shipping something real with Claude Code, Codex, Cursor, or similar — a script, an internal tool, an automation, or a documented workflow that solved a real production problem.
You speak fluent GenAI.
- Image: Magnific, GPT Image, NanoBanana, Krea, Replicate, or similar.
- Video/Motion: VEO, Seedance, Kling, Runway, Higgsfield, or similar.
- LLMs & agents: Claude, ChatGPT, Claude Code, Codex, Cursor, or similar.
You think in systems. Templates, components, variables, version control, QA loops, and repeatability are part of how you design.
You own outcomes. When the AI hallucinates, you fix it. When a request breaks the brand system, you push back with a better way. You act as if company success is your personal responsibility.
What you must have
- 4–8 years in design, motion, brand, creative production, or a similar hands-on creative role.
- A portfolio that shows full-stack capability across static design and motion/video.
- Evidence that you've built or automated at least one real creative workflow using AI — for example a script, repo, internal tool, Loom walkthrough, documented pipeline, template generator, batch production system, or automation.
- Hands-on experience with AI tools beyond manual prompting. You should be able to explain the workflow you built, the problem it solved, and how others used or could use it.
- Comfort working with GitHub or clear willingness and ability to work in a branch/PR flow.
- Strong brand judgment — you can keep AI output consistent, not just generate pretty chaos.
- Systems thinking: you do not just design one banner; you design the system that can generate 50 useful, on-brand variants.
Bonus Points
- Experience with image or video generation APIs.
- Basic scripting ability in JavaScript, Python, shell, or similar.
- Experience with no-code/low-code automation tools when they are the fastest way to solve the problem.
- Experience with DAMs, template systems, asset libraries, or creative operations tooling.
- Performance marketing creative experience across paid social, PPC, lifecycle, affiliates, or partner campaigns.
- Experience documenting workflows so other designers and marketers can use them without hand-holding.
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 is demanding work. You'll direct systems that generate thousands of variants when you used to craft one. You'll debug prompt failures and broken scripts. You'll teach non-designers to use the tools you build. You'll accept that "perfect" is sometimes the enemy of "shipped."
But you'll shape what creative looks like when production isn't the bottleneck — and build a skill set most designers don't have yet.
If you want to craft individual assets in silence, this isn't it. If you want to build the machine that creates at scale, it might be.