Compute Strategies for Generative AI - Resource Optimization and ML Platform Development

Compute Strategies for Generative AI - Resource Optimization and ML Platform Development

Learn about strategic computing resource management for large-scale AI systems in this 14-minute conference talk from Spotify's Engineering Manager ML, Avin Regmi, at the Toronto Machine Learning Series. Discover how Spotify optimizes resource allocation and minimizes wastage through their Hendrix ML Platform, which handles the training and deployment of models exceeding 70 billion parameters. Gain insights into the infrastructure and methodologies used to efficiently manage computational resources for generative AI applications at scale.