Job Description
Job Description
Join the team redefining how the world experiences design.
Hey, g’day, mabuhay, kia ora, 你好, hallo, vítejte!
Thanks for stopping by. We know job hunting can be a little time consuming and you’re probably keen to find out what’s on offer, so we’ll get straight to the point.
Where and how you can work
This role is based in Johannesburg and is a 6-month contract position.
At Canva, we believe in flexibility and trust our Canvanauts and partners to choose the balance that empowers them and their teams to achieve their goals. You’ll collaborate closely with global teammates across Research, Product, Design, and Engineering while working remotely.
What you’d be doing
As Canva scales, change continues to be part of our DNA — and nowhere is that more exciting than in AI. We’re building world-leading visual AI, and this role plays a critical part in making that happen.
You’ll sit at the intersection of design, data, and AI research, bringing world-class visual judgment into the model training and evaluation loop. In a Core Research environment, your work will materially strengthen the feedback systems that push our models toward premium, intuitive, and globally competitive performance.
At the moment, this role is focused on:
• Annotating and labeling large volumes of visual data — including images, UI layouts, compositions, graphics, and photography — with high precision
• Evaluating AI-generated outputs for aesthetic quality, composition, accessibility, usability, and brand alignment
• Developing visual scoring rubrics and structured evaluation frameworks to guide model improvement
• Identifying patterns, failure modes, and edge cases to help improve model performance
• Translating nuanced design critique into structured, actionable feedback for ML engineers and research scientists
• Using generative AI tools to prototype workflows, test hypotheses, and creatively explore model behaviors
• Contributing to the development of taxonomies for visual styles, attributes, and design patterns
• Maintaining high standards of data quality and documentation across annotation projects
You’re probably a match if
• You have strong design fundamentals — composition, color theory, typography, layout, hierarchy, and storytelling are second nature to you
• You have a portfolio that demonstrates visual excellence across graphic design, product design, UI/UX, illustration, or related domains
• You’re familiar with generative AI tools and understand how training data and annotation influence model behavior
• You can evaluate visual outputs with precision and clearly articulate structured, repeatable feedback
• You’re comfortable working in documentation-heavy, cross-functional environments
• You’re excited by the idea of shaping how AI models “see” and “create”
Bonus points if you have:
• Experience with Figma, Adobe Creative Cloud, or 3D tools
• Prior annotation, data labeling, or ML operations experience
• Basic familiarity with Python, prompt engineering, or dataset tooling
• Knowledge of accessibility guidelines (e.g. WCAG), usability heuristics, or user research
About the team
This role sits within Content Studio in the Content & Discovery Supergroup — the team focused on empowering the world through world-class content.
We combine creativity, research, and technology to ensure Canva delivers visually outstanding, accessible, and globally relevant design experiences. As Canva accelerates its AI capabilities, Content Studio plays a key role in embedding design excellence directly into model development.
You’ll collaborate with researchers, ML engineers, product managers, and designers to ensure Canva’s visual AI feels premium, human-centered, and unmistakably Canva.
Check out lifeatcanva.com for more info.
Other stuff to know
We make hiring decisions based on your experience, skills and passion, as well as how you can enhance Canva and our culture. When you apply, please tell us the pronouns you use and any reasonable adjustments you may need during the interview process.
Please note that interviews are conducted virtually.
