Research Scientist - Efficient AI (高性能大模型）
Core
Designing and advancing state-of-the-art generative AI models across multimodal domains and improving AI efficiency across the stack to power Canva's creative intelligence engine.
Role type
Research Scientist - Efficient AI (Generative Models & Systems Optimization)
Builds
Foundational generative AI models, efficient inference systems, and scalable AI features for Canva's core products.
Domain
Generative AI, Multimodal AI, Systems Optimization
Deliverable
production ML models
Required skills
foundation models (diffusion, LLMs, VLMs), low-level optimization (kernel/graph), model optimization (quantization, distillation, compression), efficient architecture design, scalable inference systems, Python, PyTorch, Transformers, DeepSpeed
Preferred skills
publications in top-tier conferences (NeurIPS, ICML, ICLR, CVPR), open-source contributions, cloud platforms expertise
Technologies
PyTorch, Transformers, Diffusers, Megatron, DeepSpeed
Responsibilities
Designing and advancing state-of-the-art generative AI models across multimodal domains, improving AI efficiency via kernel/graph optimization and model compression, collaborating cross-functionally to bring innovations into production, translating research breakthroughs into impactful features.
Seniority
Open to all experience levels (interns to experienced)