PhD Research Scientist Intern - Reinforcement Learning for Diffusion Modelling

📋 Description

• Designing and validating VLM-based evaluators for layered diffusion models.
• Turning evaluators into reward functions for RL-based generative models.
• Collaborating with research, engineering and product teams to move findings toward production.
• Contributing to Canva's layered-generation roadmap and broader research community.

🎯 Requirements

• Currently completing a PhD (ideal third year or later).
• Strong diffusion or flow-matching background with policy-gradient RL for generative models (GRPO
• Experience fine-tuning VLMs (e.g., LoRA) and designing prompts/rubrics for evaluation tasks.
• Reward modelling, preference optimization, pseudo-labelling or distillation experience.
• Ability to read a recent paper and reproduce it quickly; clear written and verbal communication.
• Enjoy collaborating with researchers and engineers on hard problems and juggling multiple threads.