The journey from training a simple model to creating a production machine learning pipeline is often seen as challenging. Typically, you have to overcome hurdles such as siloed data, inconsistent tooling and overly complex infrastructure.  The key to productionizing ML & AI is finding a solution that’s scalable, serverless and automated at every step. 

On May 21st, participants in Databrick's hands-on workshop learned how Databricks technology simplifies an organization's journey by delivering a unified, data-centric environment that uses the same platform, tools and governance for machine learning that one already uses for the rest of his or her data.

Participants were able to delve into the process of bringing an LLM application to life using cutting-edge tools. Whether an AI enthusiast eager to develop a state-of-the-art GenAI app or aiming to unlock next-gen innovation for their team or organization through GenAI use cases, this hands-on workshop promised to equip participants with practical skills, knowledge and actionable insights.

During the training, participants learned how to:

• Build and deploy your first multi-stage GenAI Agent
• Leverage context retrieval from Databricks Vector Search and Unity Catalog, governed function calling
• Access foundation LLMs offered by Amazon Bedrock, OpenAI and Databricks
• Monitor and evaluate your LLM performance using LLM judges
• Understand the concept of governed GenAI Ops and how to deploy GenAI at scale