Rad - Redesigned Relational DatabaseRad reimagines how relational databases integrate with application code. It eliminates the friction between schema definitions...
PostgreSQL Partition Pruning for Non-Partition Key QueriesThis article addresses a common PostgreSQL optimization challenge: partition pruning normally only works when filtering by the...
VLDB Volume 16: Pedreira Research PaperResearch paper from the VLDB Endowment proceedings. VLDB (Very Large Data Bases) publishes peer-reviewed research in data management and systems.
Things I want in a modern relational query languageTechnical essay exploring how functional programming concepts and modern language design could improve relational database query languages, featuring practical examples of pattern matching, sum types, and discriminated unions for SQL.
PostgreSQL for EverythingComprehensive technical article examining how PostgreSQL can replace specialized systems like Kafka, Redis, Elasticsearch, and ClickHouse for most production use cases, emphasizing simplicity and operational efficiency.
Database Detective: Minor Crimes Division on SteamAn educational puzzle game that teaches SQL querying through solving fictional crimes. A creative approach to learning database fundamentals through gameplay.
Ten years of ClickHouse in open source | ClickHouseClickHouse is an open-source analytical database that reached its 10-year milestone in June 2026. The project has grown to...
Vector Databases & Object StorageA curated reference of vector-search systems, embedded vector indexes, and object-storage-native database architectures.This...
ManyVectorManyVector provides vector and full-text search built on object storage, positioned around fast, cost-efficient, scalable search...
Amazon S3 VectorsAmazon S3 Vectors introduces vector buckets, a purpose-built S3 bucket type for storing and querying vectors.Evaluation note...
Amazon S3 Vectors User GuideThe AWS documentation entry point for understanding and using S3 Vectors.Read the S3 Vectors User Guide
Turbopuffer DocumentationThe Turbopuffer documentation and API reference.Evaluation note: Prioritise this alongside vecgo and LanceDB when comparing...
TurbopufferTurbopuffer provides vector and full-text search built on object storage, with a focus on speed, lower cost, and scale...
LanceDBLanceDB is a multimodal lakehouse for AI and an open-source embedded retrieval library for multimodal applications.Evaluation...
LanceDB GitHub RepositoryThe source repository for LanceDB, useful for inspecting its implementation and architecture.Open the LanceDB repository
vec — Go Vector Libraryvec is an embedded vector database in pure Go that aims to feel like SQLite: one database file, SIMD distance kernels, pluggable...
LanceDB GoThe Go bindings or Go implementation associated with LanceDB.Open the LanceDB Go repository
vecgoVecgo is a pure Go, embeddable hybrid vector database for production workloads, combining commit-oriented durability with HNSW...
govectorgovector is a “SQLite for Vectors” in pure Go, with embedded operation, HNSW indexing, no CGO, and a Qdrant-compatible API.Open...
Partition Finalization in Pinterest's DB Ingestion FrameworkTechnical deep-dive from Pinterest Engineering on partition finalization in their next-generation database ingestion framework, covering data pipeline architecture and stream processing.