2026/03/23 by Damian Delmas · 2 voices
Computer Science · #Advanced Database Systems and Queries #Data retrieval #Embedding #Information Retrieval and Search Behavior #Kernel (algebra) #Pipeline (software) #Query by Example #Query language #SQL #Semantic Web and Ontologies #Set (abstract data type) #cs.AI #cs.DB #cs.IR
paper · pdf · doi:10.48550/arxiv.2603.22587
openalex publication_date 2026/03/23 · arxiv published 2026/03/23 · arxiv updated 2026/03/23 · openalex created_date 2026/03/26 · openalex updated_date 2026/07/28
As AI agents become the primary consumers of retrieval APIs, there is an opportunity to expose more of the retrieval pipeline to the caller. flexvec is a retrieval kernel that exposes the embedding matrix and score array as a programmable surface, allowing arithmetic operations on both before selection. We refer to composing operations on this surface at query time as Programmatic Embedding Modulation (PEM). This paper describes a set of such operations and integrates them into a SQL interface via a query materializer that facilitates composable query primitives. On a production corpus of 240,000 chunks, three composed modulations execute in 19 ms end-to-end on a desktop CPU without approximate indexing. At one million chunks, the same operations execute in 82 ms.