Uploaded September 2026 | Updated November 2026, 3 hours ago
Microsoft AI adopted Ray to support end-to-end large reasoning model development, from pretraining and evals to RL post-training.
At Ray Summit 2026, Shaowei Su, Cheng Liu, and Pawel Beza from Microsoft AI cover Ray with heterogeneous support across Kubernetes and Slurm, scaling Ray with multi-cluster setups, topology-aware scheduling, and benchmarks around GCS, and fault tolerance for both pretraining and RL.
You'll leave with lessons from building MAI-Thinking-1 on Ray.
Liked this video? Check out other Ray Summit breakout session recordings: youtube.com/playlist?list=PLZNaJyYZRPdo
Subscribe to our YouTube channel to stay up-to-date on the future of AI: youtube.com/c/anyscale
🔗 Connect with us:
LinkedIn: linkedin.com/company/joinanyscale
X: https://x.com/anyscalecompute
Website: anyscale.com
Microsoft AI adopted Ray to support end-to-end large reasoning model development, from pretraining and evals to RL post-training.
At Ray Summit 2026, Shaowei Su, Cheng Liu, and Pawel Beza from Microsoft AI cover Ray with heterogeneous support across Kubernetes and Slurm, scaling Ray with multi-cluster setups, topology-aware scheduling, and benchmarks around GCS, and fault tolerance for both pretraining and RL.
You'll leave with lessons from building MAI-Thinking-1 on Ray.
Liked this video? Check out other Ray Summit breakout session recordings: youtube.com/playlist?list=PLZNaJyYZRPdo
Subscribe to our YouTube channel to stay up-to-date on the future of AI: youtube.com/c/anyscale
🔗 Connect with us:
LinkedIn: linkedin.com/company/joinanyscale
X: https://x.com/anyscalecompute
Website: anyscale.com










