Uploaded November 2025 | Updated September 2026, 2 weeks ago
Slides: drive.google.com/file/d/1TjWG_XEQFGPae-psRafAjHWgD8U-ceTk/view?usp=sharing
At Ray Summit 2025, Akshay Malik, Goku Mohandas, and Elizabeth Hu from Anyscale share how organizations can streamline, scale, and productionize machine learning workflows using MLOps with Ray on Anyscale.
They walk through how Ray’s unified distributed computing model—combined with Anyscale’s fully managed infrastructure—enables ML teams to move from experimentation to production with dramatic improvements in reliability, velocity, and operational simplicity. The speakers highlight best practices for building end-to-end ML pipelines, including scalable data processing, distributed training, hyperparameter tuning, batch inference, and online serving.
The session covers practical patterns for:
Simplifying MLOps through Ray-native orchestration and automation
Scaling training and inference seamlessly across clusters
Improving observability, reproducibility, and governance for ML workflows
Reducing operational overhead while increasing iteration speed
Attendees will learn how to build robust, production-grade ML systems using Ray on Anyscale—unlocking efficient, scalable MLOps for teams of any size.
Liked this video? Check out other Ray Summit breakout session recordings youtube.com/playlist?list=PLzTswPQNepXllnU0C36WtkC0dqkAoDulh
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
Slides: drive.google.com/file/d/1TjWG_XEQFGPae-psRafAjHWgD8U-ceTk/view?usp=sharing
At Ray Summit 2025, Akshay Malik, Goku Mohandas, and Elizabeth Hu from Anyscale share how organizations can streamline, scale, and productionize machine learning workflows using MLOps with Ray on Anyscale.
They walk through how Ray’s unified distributed computing model—combined with Anyscale’s fully managed infrastructure—enables ML teams to move from experimentation to production with dramatic improvements in reliability, velocity, and operational simplicity. The speakers highlight best practices for building end-to-end ML pipelines, including scalable data processing, distributed training, hyperparameter tuning, batch inference, and online serving.
The session covers practical patterns for:
Simplifying MLOps through Ray-native orchestration and automation
Scaling training and inference seamlessly across clusters
Improving observability, reproducibility, and governance for ML workflows
Reducing operational overhead while increasing iteration speed
Attendees will learn how to build robust, production-grade ML systems using Ray on Anyscale—unlocking efficient, scalable MLOps for teams of any size.
Liked this video? Check out other Ray Summit breakout session recordings youtube.com/playlist?list=PLzTswPQNepXllnU0C36WtkC0dqkAoDulh
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
