Research Scientist Graduate (High-Performance Computing (Algorithm Acceleration)- Vision AI Platform-San Jose) - 2025 Start (PhD) Join ByteDance as a Research Scientist Graduate in High-Performance Computing and Algorithm Acceleration for the Vision AI Platform in San Jose, starting in 2025. This role is ideal for PhD candidates looking to innovate in large-scale visual model development. Responsibilities Develop algorithm acceleration technologies for text-to-image/video models, including knowledge distillation, model architecture redesign (dynamic MoE routing, sparse attention), and parameter-efficient design (low-bit quantization). Lead generative model innovation, focusing on diffusion acceleration and autoregression model efficiency. Collaborate across teams to identify performance bottlenecks, optimize vision models, and enhance product capabilities. Qualifications Minimum Qualifications: PhD in Computer Science, Electrical Engineering, or related fields. Expertise in diffusion models (e.g., Stable Diffusion, DiT) and understanding of computational bottlenecks and optimization methods. Experience in at least one domain: model compression, efficient architectures, or generative alignment. Preferred Qualifications: Kaggle achievements, publications at ICML/NeurIPS/CVPR, or open-source contributions. Applicants should be able to commit to an onboarding date by the end of 2025 and clearly state their availability and graduation date in the resume. Applications are reviewed on a rolling basis, and candidates may apply for up to two positions. ByteDance is committed to inclusivity and providing accommodations for candidates with disabilities or other protected reasons. For assistance, contact us at . #J-18808-Ljbffr ByteDance