Research pillar
Multi-modal and Embodied AI cloud systems
An active systems research direction. We will add named artifacts when they are ready to share.
Our Multi-modal and Embodied AI pillar is an active systems research direction, and we will add named artifacts only when they are ready to share.
Related publications
Spotlight: Synergizing Seed Exploration and Spot GPUs for DiT RL Post-Training
Ruiqi Lai, Dakai An, Wei Gao, Ju Huang, Siran Yang, Jiamang Wang, Lin Qu, Dmitrii Ustiugov, and Wei Wang
arXiv preprint, 2026
@misc{lai2026spotlightsynergizingseedexploration, title = {Spotlight: Synergizing Seed Exploration and Spot GPUs for DiT RL Post-Training}, author = {Ruiqi Lai and Dakai An and Wei Gao and Ju Huang and Siran Yang and Jiamang Wang and Lin Qu and Dmitrii Ustiugov and Wei Wang}, year = {2026}, eprint = {2606.19004}, archiveprefix = {arXiv}, primaryclass = {cs.DC}, url = {https://arxiv.org/abs/2606.19004} }CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference
Yulin Zou, Yan Chen, Wenyan Chen, JooYoung Park, Shivaraman Nitin, Luo Tao, Francisco Romero, and Dmitrii Ustiugov
arXiv preprint, 2026
arXiv2 citations@misc{zou2026codecsightleveragingvideocodec, title = {CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference}, author = {Yulin Zou and Yan Chen and Wenyan Chen and JooYoung Park and Shivaraman Nitin and Luo Tao and Francisco Romero and Dmitrii Ustiugov}, year = {2026}, eprint = {2604.06036}, archiveprefix = {arXiv}, primaryclass = {cs.DC}, url = {https://arxiv.org/abs/2604.06036} }