databricks-ai-functions

What problem does it solve?

Databricks AI Functions provide built-in AI capabilities within SQL and PySpark pipelines, eliminating the need to manage external model endpoints or custom inference deployments.

Core Features & Use Cases

• Task-specific functions for common analytics tasks (ai_classify, ai_extract, ai_summarize, ai_translate, ai_parse_document, ai_query, ai_forecast, ai_analyze_sentiment, ai_mask, ai_gen, ai_similarity) that operate directly on table columns and scale in batch workloads.
• Document ingestion and RAG-style patterns: parse PDFs and office documents with ai_parse_document, then enrich with targeted extractions or structured JSON via ai_query when necessary.
• End-to-end data workflows: apply AI functions to support ticket analysis, invoice processing, and time-series forecasting without managing model endpoints.

Quick Start

Run an example: classify a snippet of text using ai_classify and extract fields with ai_extract.