Empowering Generative AI with Databricks and AMD

This session explores the essential role of GPU hardware in the pretraining and fine-tuning of generative AI models, highlighting the shift from a single-vendor-dominated industry with limited GPU availability. We'll discuss AMD's contributions with advanced GPU hardware and ROCm software that are transforming the ability to train models at scale. Emphasis will be on leveraging AMD GPUs for processing private datasets, using Delta Sharing and Unity Catalog for seamless environment transitions, and serverless infrastructures for inference to aid monetization. Participants will gain insights on how Databricks & AMD make Generative AI more accessible and cost-effective, all while ensuring robust governance and data privacy. We'll also cover comprehensive best practices for the complete training and deployment lifecycle of generative AI models.