Video Foundation Models Scientist — Multimodal AI

Cantina; San Francisco (CA); Vor Ort; USD 200.000 - 320.000; Vollzeit

Benefits dieser Stelle

Competitive salary

Company equity

Medical insurance

Dental insurance

Vision insurance

Paid time off

Parental leave

Fertility support

401(k) plan

Lunch provided

One Medical membership

Zusammenfassung

Cantina Labs is building a core team to advance native video and multimodal foundation models for scalable generation and understanding. You will shape architecture, data, training, and evaluation across the full model lifecycle while collaborating with researchers, engineers, and product teams to drive impactful, real-world capabilities.

The role focuses on pre-training, continued training, and post-training for high-quality, controllable generation with opportunities for publications and

Qualifikationen

• Experience with diffusion models, flow matching, DiTs, or related generative methods.
• Hands-on experience training large-scale image/video or multimodal models.
• Proven track record through publications, open-source contributions, or production impact.
• Ability to own ambiguous research problems and collaborate in a team.
• Specialized depth across foundation model lifecycle: architecture, data, post-training, deployment.

Aufgaben

• Research, develop, and scale native video and multimodal foundation models.
• Explore architectures, training objectives, and conditioning for video generation.
• Build data curation, distributed training, evaluation pipelines for quality control.
• Design experiments to understand scaling, generation quality, controllability, and efficiency.
• Collaborate with researchers, engineers, and product teams to shape roadmaps.
• Contribute to research publications and open-source releases when appropriate.

Kenntnisse

Generative modeling

Diffusion models

Video generation

Multimodal models

Distributed training

Research publications

Jobbeschreibung

Cantina Labs is building a core team to advance native video and multimodal foundation models for scalable generation and understanding. You will shape architecture, data, training, and evaluation across the full model lifecycle while collaborating with researchers, engineers, and product teams to drive impactful, real-world capabilities.

The role focuses on pre-training, continued training, and post-training for high-quality, controllable generation with opportunities for publications and