Ajay Kulkarni and Mike Freedman on Vector Search, Multimodal Data, and AI/ML capabilities of Postgres.
Subscribe: Apple • Spotify • Overcast • Pocket Casts • AntennaPod • Podcast Addict • Amazon • RSS.
Ajay Kulkarni and Mike Freedman are the co-founders of Timescale, a startup that provides an enhanced version of PostgreSQL optimized for time-series analytics, AI applications, and scalable relational workloads. We discuss the evolution of Postgres from a traditional relational database to a versatile platform that can handle various workloads, including vector search for generative AI applications. The scalability, performance, and cost-efficiency of Postgres make it a compelling alternative to using multiple specialized databases.
Relevant posts from Timescale:
- How We Made PostgreSQL as Fast as Pinecone for Vector Data
- Why PostgreSQL Is the Bedrock for the Future of Data
Interview highlights – key sections from the video version:
- Introducing pgvectorscale: A New Vector Search Capability for Postgres
- The Evolution of Postgres and the Need for pgvectorscale
- When and How to Use pgvectorscale
- Exploring Hybrid Search and Data Integration in Postgres
- Scaling Vector Search: pgvectorscale’s Capabilities and Use Cases
- Addressing Scalability Concerns: Timescale Cloud and Horizontal Expansion
- Real-Time Index Updates and Incremental Indexing in pgvectorscale
- Supported Vector Indices and Statistical Binary Quantization
- Leveraging DBMS Features: Governance, Access Control, and Beyond
- Ease of Use and Integration: Postgres vs. Specialized Vector Search Systems
- Streaming Post-Filtering: Addressing Scarce Data Challenges
- Beyond Vector Search: Exploring New File Formats (Lance and Nimble)
- Postgres for Everything: Evidence and Examples
- Embedded Databases: The Rise of DuckDB, SQLite, and Their Relationship to Postgres
- Workloads Where Postgres Might Not Be the Best Fit
- Looking Ahead: Next Innovations in Generative AI and Postgres
- Postgres and Multimodal Data: An “Everything Modal” Database
- Key Postgres Events and Resources for the Community
Related content:
- A video version of this conversation is available on our YouTube channel.
- Previous episodes pertaining to vector search can be found HERE.
- Choosing the Right Vector Search System
- Is Your Data Strategy Ready for Generative AI?
- Distributed Systems: Not Just Scale
- Emil Eifrem → The Future of Graph Databases
- Semih Salihoglu → The Intersection of LLMs, Knowledge Graphs, and Query Generation
- Joao Moura → Unleashing the Power of AI Agents
If you enjoyed this episode, please support our work by encouraging your friends and colleagues to subscribe to our newsletter: