kadence ยท 3 days ago
Research Scientist - Video Generation
Kadence is a well-funded AI research lab focused on building state-of-the-art video generation models for cinematic applications. They are seeking a Research Scientist with expertise in video generation and diffusion models to lead research efforts and transition findings into production systems.
Responsibilities
Conduct original research in video generation, diffusion models, and generative modeling
Design, implement, and train large-scale video diffusion models
Improve temporal consistency, motion realism, and long-horizon coherence in generated video
Collaborate closely with engineering and production teams to transition research into deployed systems
Publish and present research at top-tier conferences when appropriate
Stay current with the latest advances in generative modeling and video synthesis
Qualification
Required
PhD in Machine Learning, Computer Vision, AI, or a related field (or equivalent research experience)
First-author publications in video generation, diffusion models, or closely related domains
Relevant venues include: NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, SIGGRAPH, SIGGRAPH Asia
Strong background in: Diffusion models and generative modeling
Video representation learning and temporal modeling
Deep learning frameworks (e.g., PyTorch)
Proven ability to translate research ideas into working, scalable systems
Preferred
Experience working with large-scale video datasets
Prior work in image/video synthesis for media, film, or creative tools
Experience moving research into production environments
Open-source contributions or widely cited research
Benefits
Competitive salary and meaningful equity
Opportunity to work on models deployed in major film productions
High-impact research role with significant technical ownership
Collaboration with a small, elite research team
Strong financial backing and long-term runway
Company
kadence
Kadence - We place elite AI talent at the intersection of science & engineering, from PHD to Executive.
Funding
Current Stage
Early StageCompany data provided by crunchbase