Lakehouse to LLM: Solve Business Tasks with LLMs Reliably and at Scale

Can we process large amounts of data with LLMs for workflow automation and avoid errors and hallucinations? We will dive deep into running AI models at scale to automate multiple back-office tasks, breaking out from the one-on-one chat experience with LLMs onto massive-scale AI-powered databases and workflows. This session will show how you can empower your team with LLM-powered workflows that process multimodal data from Databricks, allow for continuous improvement and fine-tuning, and enable you to specialize AI models on your business task. During this session, we will use models like GPT-4 Vision, Claude, and Llama to automate document processing, web page auditing, image recognition, and financial analysis tasks at an enterprise-scale level. This session is best suited for practitioners who are ready to implement LLMs into complex human-led processes for increased automation via specialized co-pilots. It will be a fully UI-led talk with data analysis but no code.