该岗位现面向所有经验阶段的候选人开放，包括社会招聘、2025年及2026年应届毕业生，同时开放实习生岗位。工作地点为北京。欢迎申请，期待你的加入！

Notice: This position is open to candidates at all experience levels, including experienced candidates, 2025 and 2026 graduates, as well as internship opportunities. The role is based in Beijing. We welcome your application and look forward to having you on board!

At Canva, we're building a future powered by AI that's as magical as it is impactful. As a Research Scientist at Canva, you'll be responsible for advancing the future of AI by experimenting with cutting-edge techniques, as well as improving models for real-world quality and performance.

About the role:

You will focus on building agentic AI models with the latest multimodal generation models to revolutionize productivity software. You will lead efforts to bring the best existing AI models to hundreds of millions of active users. You will be able to coordinate with top talent with deep domain expertise and great passion.

At the moment, this role is focused on:

• Applying Research: Own application-focused research topics, rapidly track and validate the latest industry advancements. Turn outcomes from internal or external fundamental generative models into high-impact applications with innovations

• Designing New Frontier Tasks: As AI models have unlocked mind-blowing scenarios like DeepResearch and CUA agents, you will conduct research on expanding what the latest multimodal AI models can generate by defining new challenging tasks that better align with real user requirements. Collaborate with ML Engineers to identify emerging technical trends and continuously explore new application directions and scenarios

• SFT + Post-Training: Abstract and address critical and challenging user needs with AI models. Perform post-training of fundamental models to address various downstream applications like ensuring character consistency, accurate glyph generation, and enhanced visual aesthetics

You’re probably a match if you have:

• Rich experience in shipping the latest generative models like FLUX based on Diffusion Transformer, Qwen3-Coder, Qwen2.5-VL, and DeepSeek- R1.

• Ability to translate advancements in the latest generative models into robust, production-ready implementations

• Experience with deploying diffusion models and LLMs, and improving efficiency either through distillation, speculative sampling strategies, or quantization.

• Collaborative mindset — you work exceptionally well with engineers, product teams, and other researchers. Partner closely with ML Engineers to accelerate model inference and integrate them into the product ecosystem

• Ability to share your knowledge internally through documentation and cross-team collaborations, and externally through papers or talks where appropriate

• Proficiency in Python, PyTorch, and cloud computing platforms, ensuring fast model deployment