25 min
AI Architecture
Open Source
Founder Stories
Episode 22: Taming AI Complexity with Head of Engineering at Anyscale
25 min
Jaikumar Ganesha (JK), Head of Engineering at Anyscale, breaks down modern AI infrastructure challenges and how Ray (the open-source compute framework) is powering production-grade AI at scale. Covers the 'AI complexity wall,' distributed training, and includes a live demo of building a multimodal AI app.
Timestamps
00:00Intro & Speaker Welcome01:00AI Complexity Wall03:00Why 90% of AI projects never reach production04:30What is Ray? Introduction06:00Key Trends: Multimodal, Agentic AI & Post-Training09:00How Ray supports complex ML use cases12:00The Modern AI/ML Stack15:00Ray Data for Multimodal Workloads17:30Pipeline Example: Resize, Segment, Classify20:00Ray + Anyscale: Making Production Easy22:30Demo: Developer Workflow & Auto-Scaling24:00Recap & Try Ray Today