This role involves contributing to the development of cutting-edge multimodal AI foundational models at Microsoft AI. Responsibilities include algorithm development, model architecture design, experimentation, measurement, evaluation, dataset innovation, and pipeline improvement. The ideal candidate will possess expertise in multimodal research, a strong publication record, and experience with large-scale distributed systems. The work emphasizes a data-driven approach, collaboration, and responsible AI practices. The team aims to improve training and deployment efficiency and innovate across prototypes and products. Collaboration with infrastructure, data engineering, pre-training, post-training, and product feedback teams is crucial.