Databricks · Databricks Generative AI Engineer Associate
Building the Chain
Databricks Generative AI Engineer Associate · Lesson 10 of 17 · 1:15:40
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the Databricks Certified Generative AI Engineer Associate exam. In this lesson: a custom pyfunc model (PythonModel, load_context once, predict per request), pre-processing and post-processing around the chain, its parts (retriever, prompt template, chat model, output parser), what to log (flavor, dependencies, input example, signature, resources), and registering in Unity Catalog with versions and aliases. Includes three original practice questions.
Preview the topics
- Copperline’s pyfunc model opens a new connection to the AI Search index inside its predict method, and every request is slow. Where should that connection be opened?
- Copperline’s chain works in the notebook, but the serving endpoint fails at start-up, with an import error for a PDF parsing package. What was left out when the model was logged?
- Copperline registered version 7 of its assistant in Unity Catalog. Production code loads the model by its champion alias. How should the team promote version 7?