Tobias Kreiman

Papers

Total Citations

H-Index

About

Tobias Kreiman is a rising star in robotics and machine learning, whose work is helping to democratize generalist robot intelligence. His primary research focuses on developing open-source, large-scale policies that can transform how robots learn, moving away from task-specific training toward versatile, pretrained models. Kreiman’s most notable contribution is **Octo**, an open-source generalist robot policy that has already garnered significant attention, with its 2024 publication accumulating over 66 citations. This work demonstrates that large policies pretrained on diverse robot datasets can be fine-tuned with minimal in-domain data while still generalizing broadly across tasks and environments. By releasing Octo to the community, Kreiman is accelerating progress in robotic learning, enabling researchers and developers to build on a shared foundation rather than starting from scratch. His impact lies in making advanced robot learning more accessible, scalable, and practical—a crucial step toward robots that can adapt to the real world. For students and researchers, Kreiman’s work represents a pivotal shift toward open, reusable, and generalist approaches in embodied AI.