Cascading retrieval with multi-vector representations — ConstBERT model is now in open source - Read the blog
Publications

Pinecone Research

Pinecone's scientists and engineers work on cutting-edge research at the intersection of information retrieval, model training, and efficient system architectures. These insights form the foundation of the company's state-of-the-art vector database, including its algorithms and system architecture, and also appear in academic conferences, scholarly publications, and other venues.

Featured
ECIR 2025 (Best Short Paper, Honourable Mention)

Efficient Constant-Space Multi-Vector Retrieval

Sean MacAvaney, Antonio Mallia, Nicola Tonellotto

Start building knowledgeable AI today

Create your first index for free, then pay as you go when you're ready to scale.