Leading real-time analytics database now stores and indexes vector embeddings, enabling efficient AI-powered personalization, recommendations and anomaly detection at scale
Rockset, the real-time analytics database built for the cloud, today announced native support for vector embeddings, enabling organizations to build high-performance vector search applications at scale, in the cloud. By extending its real-time SQL-based search and analytics capabilities, Rockset now allows developers to combine vector search with filtering and aggregations to enhance the search experience and optimize relevance by enabling hybrid search.
Vector search has gained rapid momentum as more applications employ machine learning (ML) and artificial intelligence (AI) to power voice assistants, chatbots, anomaly detection, recommendation and personalization engines—all of which are based on vector embeddings at their core. Rockset delivers fast, efficient search, aggregations and joins on real-time data at massive scale by using a Converged Index™ stored on RocksDB. Vector databases, such as Milvus, Pinecone, Weaviate and other popular alternatives like Elasticsearch, store and index vectors to make vector search efficient. With this release, Rockset provides a more powerful alternative that combines vector operations with the ability to filter on metadata, do keyword search and join vector similarity scores with other data to create richer, more relevant ML and AI powered experiences in real-time.
“By extending our existing real-time search and analytics capabilities into vector search, we give AI/ML developers access to real-time data and fast queries with a fully managed cloud service,” said Rockset co-founder and CEO Venkat Venkataramani. “We now enable hybrid metadata filtering, keyword search and vector search, simply using SQL. Combining this ease of use with our compute efficiency in the cloud makes AI/ML a lot more accessible for every organization.”
With the new release, Rockset supports vector operations along with the following benefits:
- Real-time data: High velocity real-time indexing with support for updates
- Fast search: Combine vector search, keyword search and metadata filtering for fast, more efficient results
- Hybrid search and analytics: Join vector similarity scores with other data to create richer, more relevant experiences, using SQL
- Fully managed cloud service: A horizontally scalable, highly available cloud-native database with compute-storage and compute-compute separation for cost-efficient scaling in the cloud
Rockset can also be used with OpenAI’s Embeddings API to generate, index and query language embeddings. In addition, the company announced an integration with the Feast Feature Store to streamline the management of vector embeddings across multiple data sources and frequently updated models. With the ability to scale to billions of vectors and serve thousands of queries per second, Rockset enables real-time machine learning efficiently at scale.
Learn more about the release by joining the webinar: How to build Real-Time Machine Learning at Scale.
Visit AITechPark for cutting-edge Tech Trends around AI, ML, Cybersecurity, along with AITech News, and timely updates from industry professionals!