Uploaded May 2025 | Updated September 2026, 1 week ago
Attentive is redefining digital marketing for over 500 million consumers with AI-powered personalization, powered by Ray and Anyscale.
Learn how they optimized training workflows, deployed large-scale recommendation systems, and tackled infrastructure challenges to deliver smarter, faster, and more relevant shopping experiences.
Whether you're building large-scale ML systems, working on personalization engines, or looking to optimize your ML infrastructure, this talk provides battle-tested strategies and real-world lessons learned from achieving results at massive scale. Perfect for ML engineers, data scientists, and technical leaders looking to understand how to deliver game-changing personalization results.
00:00 - 03:26 - How does Attentive solve personalization
03:27 - 06:08 - Prior model architecture & challenges
06:09 - 10:12 - Two tower model overview
10:13 - 14:12 - Solving for Scale with Ray + Anyscale
14:13 - 18:58 - PyTorch to Ray migration
18:59 - 22:00 - Model Results
22:01 - 24:39 - Platform Adoption Metrics
--
Liked this video?
Check out the other presentations: youtube.com/playlist?list=PLzTswPQNepXlWVHzNpyWxyvyzeOV-5-uA
--
🔗 Connect with us:
- Subscribe to our YouTube channel: youtube.com/@anyscale
- Twitter: https://x.com/anyscalecompute
- LinkedIn: linkedin.com/company/joinanyscale
- Website: anyscale.com
Attentive is redefining digital marketing for over 500 million consumers with AI-powered personalization, powered by Ray and Anyscale.
Learn how they optimized training workflows, deployed large-scale recommendation systems, and tackled infrastructure challenges to deliver smarter, faster, and more relevant shopping experiences.
Whether you're building large-scale ML systems, working on personalization engines, or looking to optimize your ML infrastructure, this talk provides battle-tested strategies and real-world lessons learned from achieving results at massive scale. Perfect for ML engineers, data scientists, and technical leaders looking to understand how to deliver game-changing personalization results.
00:00 - 03:26 - How does Attentive solve personalization
03:27 - 06:08 - Prior model architecture & challenges
06:09 - 10:12 - Two tower model overview
10:13 - 14:12 - Solving for Scale with Ray + Anyscale
14:13 - 18:58 - PyTorch to Ray migration
18:59 - 22:00 - Model Results
22:01 - 24:39 - Platform Adoption Metrics
--
Liked this video?
Check out the other presentations: youtube.com/playlist?list=PLzTswPQNepXlWVHzNpyWxyvyzeOV-5-uA
--
🔗 Connect with us:
- Subscribe to our YouTube channel: youtube.com/@anyscale
- Twitter: https://x.com/anyscalecompute
- LinkedIn: linkedin.com/company/joinanyscale
- Website: anyscale.com










