Retrieval Augmented Generation (RAG) is a technique that grounds a model's responses in retrieved external data rather than only its training. The GenAI Engineer exam centers on building RAG applications on Databricks.
Retrieval Augmented Generation (RAG) is a technique that grounds a model's responses in retrieved external data rather than only its training.
Databricks Vector Search is a serverless vector database that indexes embeddings for similarity retrieval in RAG applications.
An embedding is a numeric vector representation of text that captures semantic meaning for similarity search.
Chunking is the process of splitting documents into smaller passages so they can be embedded and retrieved effectively.
Mosaic AI is the Databricks suite for building, serving, and evaluating AI and generative AI applications.
Databricks Foundation Model APIs provide access to hosted large language models for inference from within Databricks.
LangChain is an open-source framework for composing LLM calls, tools, and retrievers into chains and agents.
An agent is a generative AI component that plans and calls tools over multiple steps to complete a task.
MLflow is an open-source platform, deeply integrated with Databricks, for tracking, logging, and registering models and AI applications.
Databricks Model Serving is a service that deploys models and AI applications behind scalable REST endpoints.
Unity Catalog is the Databricks governance layer that manages access, lineage, and permissions for data, models, and AI assets.
Governance guardrails are safety filters and policies that constrain generative AI inputs and outputs, such as blocking unsafe content or masking PII.
LLM-as-a-judge is an evaluation approach that uses a language model to score the quality of generative AI outputs.
An inference table is a Databricks table that automatically logs the requests and responses of a served model for monitoring.