Research Scientist - Multimodal Interaction and Model - Pre-Training

Position: Research Scientist - Multimodal Interaction and World Model - Pre-Training
About Seed Team:
Established in 2023, the Byte Dance Seed team is dedicated to discovering new approaches to general intelligence, and pursuing the edge of intelligence. Our research spans large language models, Gen Media, AI for Science, Robotics and more. With a long-term vision and determination in AI, the Byte Dance Seed team remains committed to foundational research. We aim to become a world-class AI research team that drives real technological progress and delivers societal benefits.

With labs across China, Singapore, and the U.S., our team has already released industry-leading general-purpose large models and advanced multimodal capabilities, powering over 50 real-world applications - including Doubao, Coze, and Jimeng.

- Drive research and engineering to advance models that enhance understanding of multimodal data and enlarge reasoning capabilities.

- Design novel architectures that optimize both the model's performance and efficiency.

- Establish scaling laws, design and conduct systematic ablations that result in transferrable conclusions.

Minimum Qualifications:

- Expertise in Transformers (Dense and MoE) and familiar with how to scale Transformers on GPUs or TPUs.

- Rich hands-on experience in PyTorch/JAX and distributed training framework.

- Familiar with state-of-the-art techniques for preparing multimodal training data.