Research Intern – Video Generation Models

About the role

Cantina Labs is expanding its research lab in Singapore and seeks a research intern to contribute to the next generation of video generation models. The three‑month onsite internship offers ownership of a well‑defined research problem, mentorship from senior researchers, and the possibility of a conference publication.

Key responsibilities

• Research and develop distillation methods for large‑scale diffusion and flow‑based video generation models, including guidance and adversarial distillation.
• Explore techniques to reduce inference cost while preserving or improving generation quality, and design rigorous evaluations and large‑scale experiments on generative video models.

Required profile

• Currently pursuing a PhD or final‑year master’s in computer science, machine learning, computer vision, or a related field.
• Research experience in generative modeling, computer vision, multimodal learning, or video generation.

Required skills

• Proficiency in Python.
• Hands‑on experience with PyTorch or JAX.
• Experience with diffusion models, flow‑based models, and model distillation.
• Knowledge of reinforcement learning or preference‑based optimization techniques.
• Familiarity with video, image, audio, or other multimodal data.

What we offer

• Competitive monthly stipend with visa and travel support for eligible international candidates.
• Housing assistance for qualifying interns in Singapore and provision of equipment and compute resources.
• Mentorship from senior researchers, clear project milestones, and potential first‑author conference publication.