Student Researcher: Long-Range Video Generation

ByteDance; San Jose (CA); On-site; USD 41,000 - 69,000; Part time

Job summary

ByteDance Seed is seeking an intern to contribute to foundational models for visual generation within the Seed team based in San Jose. You will collaborate on research for multimodal generative models, with emphasis on video and temporal sequences, under the Technology group.

The role targets PhD students with strong publications and hands-on experience in deep learning, diffusion/transformer video models, and large-scale datasets.

Qualifications

• Currently pursuing a PhD in Computer Vision, Machine Learning, or a related field.
• Research experience in generative modeling, especially for video, motion, or temporal sequences.
• First-author publications in CVPR, ICCV, ECCV, NeurIPS, ICLR, or ICML.
• Proficiency in deep learning frameworks and experience with large-scale video datasets.

Responsibilities

• Develop scalable architectures for long-range video generation with consistent motion, identity, and layout.
• Explore hierarchical or recurrent latent structures to support generation across long temporal spans.
• Address challenges in temporal drift, motion collapse, and high-frequency detail retention.
• Investigate autoregressive or chunked generation strategies that balance quality and memory.
• Design evaluation protocols for long video quality (e.g., realism, consistency, semantic continuity).

Skills

Video generation research

GenAI research

Deep learning

Education

PhD student in Computer Vision or ML

Tools

PyTorch

Large-scale video datasets

Job description

ByteDance Seed is seeking an intern to contribute to foundational models for visual generation within the Seed team based in San Jose. You will collaborate on research for multimodal generative models, with emphasis on video and temporal sequences, under the Technology group.

The role targets PhD students with strong publications and hands-on experience in deep learning, diffusion/transformer video models, and large-scale datasets.