AI Research Scientist 大模型研究科学家
Core
Advancing frontier AI research by pre-training and post-training multi-modal generation models (text-to-image/video/3D) to improve real-world quality and performance for productivity software.
Role type
AI Research Scientist (Large Language Models & Generative AI)
Builds
Frontier multi-modal generation models (diffusion, auto-regressive) for Canva's productivity software used by hundreds of millions of users.
Domain
Generative AI, Multimodal Models, Large-Scale Deep Learning
Deliverable
production ML models
Required skills
Large-scale model pre-training, Diffusion models, Auto-regressive models, RL algorithms, Synthetic data generation, High-end GPU cluster management, Python, PyTorch, Transformers, DeepSpeed, Megatron
Preferred skills
Peer-reviewed publications in top-tier conferences (CVPR/ICCV/NeurIPS/ICML/ICLR), Open-source contributions, Cross-team collaboration
Technologies
PyTorch, Diffusers, Transformers, Megatron, DeepSpeed, High-end GPUs
Responsibilities
Determine and lead research initiatives aligned with frontier multimodal AI trends, Design and run experiments to validate hypotheses on high-end GPUs, Interpret experimental results and translate them into actionable insights, Create novel generative model architectures and training methodologies
Seniority
All experience levels (Entry to Senior, Interns welcome)
