Databricks' KARL agent uses reinforcement learning to generalize across six enterprise search behaviors — the problem that breaks most RAG pipelines.

Students had to compile a portfolio and discuss it in an interview to demonstrate their understanding and ability to apply the knowledge.