What You'll Do
Lead the development of robust, scalable tools and APIs that allow large language models to interact reliably with design functionality. Define the architectural patterns for tool interfaces, payload structures, and intent-driven communication, ensuring consistency across platforms.
Design and maintain evaluation systems that measure accuracy and performance of AI tool usage, supporting both rapid proxy-based assessments and real-world production monitoring. Ensure these frameworks evolve as new AI platforms and features are added.
Shape the strategic direction of Canva’s agent architecture by determining optimal placement of intelligence—balancing external LLMs with internally hosted agents. Build orchestration layers that empower third-party systems to invoke design operations at scale.
Develop observability solutions that provide insight into how AI assistants consume design tools in production. Use data to identify failure modes, set quality standards, and drive iterative improvements. Automate complex workflows such as multi-step marketing campaign generation.
Collaborate across engineering teams and external AI partners to solve emerging challenges. Influence integration strategies with leading AI platforms and help guide internal teams on best practices for AI consumption.
Requirements
- Proven experience deploying LLM-powered systems into production, with measurable outcomes such as improved tool-call accuracy, reduced error rates, or lower latency
- History of identifying unstructured problems in AI-tool interaction and driving end-to-end solutions across teams
- Ownership of evaluation pipelines—from metric design to infrastructure—used to inform architectural decisions at scale
- Established technical standards in tool design, API contracts, or observability that have been adopted across teams
- Ability to translate shifts in external AI capabilities into strategic adjustments within internal systems
- Proficiency in Python and ML frameworks, with strong working knowledge of TypeScript/Node.js and cloud environments such as AWS or Cloudflare Workers
- Experience influencing technical direction without formal authority, particularly in cross-functional or partner-facing contexts
- Thrives in fast-moving, ambiguous environments and consistently delivers clarity and progress
Technical Stack
Python, machine learning frameworks, TypeScript, Node.js, Cloudflare Workers, AWS
Work Mode
This role supports remote work within Australia, enabling flexible collaboration across distributed teams while contributing to a globally used design platform.
