From the Data Exchange podcast, we present recent conversations with creators of popular vector databases. The popularity of vector databases and vector search has been increasing among technical teams. This is mainly due to the widespread use of dense vector representations of data, made possible by advances in neural networks. Furthermore, the decision of technology giants like Yahoo and Facebook to open source their core systems for vector search (Yahoo Vespa, Facebook Faiss) has further fueled this interest. The trend is also reflected in the growing number of vector database startups that have raised nearly $200 million in funding, leading to the emergence of new enterprise solution providers and advocates.
Vector Database IndexWe compare vector databases and libraries using an index that measures popularity. For this inaugural edition, we focus on specialized systems and include only one general search engine – Elasticsearch, which incorporated vector search through Apache Lucene’s new ANN capabilities.
An open source, production grade vector search engineBob van Luijt, is CEO of SeMI Technologies, the company behind the popular vector search engine Weaviate. Weaviate has integrations with OpenAI, Huggingface, Cohere, and other popular tools.
A new storage engine for vectorsRam Sriharsha is VP of Engineering and R&D at Pinecone, a startup that offers a fully managed vector database (not just an index).
A Cloud Native Vector Database Management SystemFrank Liu is Director of Operations & ML Architect at Zilliz, the company behind Milvus, an open source vector database.
[Image: Embeddings by Ben Lorica; original photos sourced on Infogram.]