Description
Vector databases such as ElasticSearch and Pinecone offer fast ingestion and querying on vector embeddings with ANNs. However, they typically do not decouple compute and storage, making them hard to integrate in production data stacks. Because data storage in these databases is expensive and not easily accessible, data teams typically maintain ETL pipelines to offload historical embedding data to blob stores. When that data needs to be queried, they get loaded back into the vector database in another ETL process. This is reminiscent of loading data from OLTP database to cloud storage, then loading said data into an OLAP warehouse for offline analytics. Recently, “lakehouse” offerings allow direct OLAP querying on cloud storage, removing the need for the second ETL step. The same could be done for embedding data. While embedding storage in blob stores cannot satisfy the high TPS requirements in online settings, we argue it’s sufficient for offline analytics use cases like slicing and dicing data based on embedding clusters. Instead of loading the embedding data back into the vector database for offline analytics, we propose direct processing on embeddings stored in Parquet files in…
Description from YouTube. Full content on the video page.
Topics
More from Databricks
NewsParallel Coding Agents with Lakebase | Claude Code + GitHub Actions
This video demonstrates how to run multiple coding agents in parallel by combining git worktrees, GitHub Actions, and Lakebase database branching. It shows how each agent automatically receives an isolated database branch for safe experimentation and schema migrations, followed by dedicated preview environments for pull requests.
EventsHow Enterprises Govern AI Agents Across Multiple Models
Databricks announced the general availability of the Unity AI gateway to provide centralized multi-model governance, cost controls, and end-to-end observability for enterprise AI agents. Panelists discussed how coding agents and harnesses are evolving beyond programming into long-running operations, personal software development, and automated organizational workflows.
EventsDemo: Building a Governed AI Agent with Unity AI Gateway
This video demonstrates how to build, update, and govern a store operations AI agent using Databricks Agent Bricks and the Unity AI Gateway. The tutorial highlights integrating custom Model Context Protocol servers, recording execution traces with MLflow, and enforcing security policies and budget controls.
NewsHow ModMed Transforms Healthcare AI and Agentic Workflows with Databricks
ModMed uses the Databricks Lakehouse platform and Unity Catalog to build secure AI-enabled healthcare applications and agentic workflows. The integration of these tools allows both technical and non-technical users to access near real-time data insights and solve complex problems efficiently.
NewsTeach AI how your business actually runs
Model intelligence is no longer the bottleneck for enterprise AI adoption because modern frontier models easily handle complex reasoning tasks. Business value requires providing these models with specific organizational context and metadata about internal processes to create a competitive advantage.