Model Research Scientist

Our client is building next-generation video generation models that simulate environments and how scenes evolve over time. The team is focused on improving how generative video models behave over long horizons, making them more controllable, consistent, and useful for interactive systems.

The role focuses on applying reinforcement learning and robotics-style post-training techniques to video diffusion models to improve cont rollability and long-horizon behaviour in generative video systems. (This is not a robotics role, the work is focused on video generation models.)

What you’ll work on:

• Applying reinforcement learning, RLHF, and robotics-style training methods to improve video diffusion and generative video models
• Designing reward signals, training objectives, and evaluation metrics that improve temporal consistency and long-horizon generation
• Building systems that generate simulated environments and multi-step tasks for training video models
• Improving training pipelines and evaluation frameworks used to measure model capability and behaviour
• Working closely with researchers and engineers to iterate quickly on new training approaches

What we’re looking for:

• Around 1+ year of experience outside of a Master’s or PhD working with reinforcement learning or related training methods - on video generation
• Experience applying robotics-style training techniques such as imitation learning, reward modeling, offline RL, or RLHF
• Experience working with video generation models, diffusion models, or other generative models
• Strong ML/RL fundamentals including reward design, policy optimisation, and evaluation
• Ability to run experiments end-to-end and iterate quickly
• Strong programming skills in Python and modern deep learning frameworks such as Py Torch

Bonus:

• Publications or open-source contributions in machine learning, reinforcement learning, or generative modelling
• Experience working with large-scale model training or distributed ML systems

Our client is an equal opportunity employer and welcomes applicants from all backgrounds. All qualified candidates will receive consideration for employment without regard to race, colour, religion, sex, sexual orientation, gender identity, national origin, disability, or veteran status.