Sudhir Hasbe on vector search, analytics, lakehouses, knowledge graphs, and more.
Subscribe: Apple • Spotify • Overcast • Google • AntennaPod • Podcast Addict • Amazon • RSS.
Sudhir Hasbe, Chief Product Officer at Neo4j, and a longtime technical and product leader in the data management space. We discussed several important trends including:
- Vector Search and Vector Databases
- Analytics (including popular open source projects like DuckDB, Clickhouse, and Postgres)
- Lakehouses
- Knowledge Graphs
- Multimodal systems and HTAP
-
❛We’re all in data space, and we’re all capable of building the same thing. Just look at the last six months. How many existing systems were able to add vector support to their databases? Almost every major database provider has already done it. What I believe is that it’s much easier for the core database companies to add vector indexes and vector searches within themselves than it is for a dedicated vector store to integrate all the capabilities that different systems offer. In eternity, they could probably achieve anything. I’m pretty confident that databases like Postgres, and many others, can incorporate all the necessary capabilities. So, I don’t see that as a challenge. For instance, we were able to add vector support to our database in less than three months and then rolled it out to all our customers. And we’ll only improve from here, blending capabilities between knowledge graphs and vector searches, both explicit and implicit relationships. We can merge these features for various use cases. So, what I’m getting at is that it’s easier for existing database companies to add vectors than for vector databases to integrate all the functionalities of a relational database or a NoSQL system like Mongo, or search systems like Elastic. All of those are commendable in their own right.❜
– Sudhir Hasbe, Chief Product Officer at Neo4j.
Interview highlights – key sections from the video version:
- Vector Search and Vector Databases
- Are standalone vector databases going to be necessary in the long run?
- Open source analytic databases: DuckDB, Clickhouse, Postgres, and others.
- The rise of Lakehouses
- Knowledge graphs and retrieval-augmented generation
- Multimodal systems and HTAP
Related content:
- A video version of this conversation is available on our YouTube channel.
- Vector Database Primer
- Emil Eifrem: The Future of Graph Databases
- Jesse Anderson, Evan Chan, and Ben Lorica: 2023 Trends in Data Engineering and Infrastructure
- Navigating the Intricacies of LLM Inference & Serving
- Specialized Hardware for AI: Rethinking Assumptions and Implications for the Future
- Brian Raymond: ETL for LLMs
- Jerry Liu: An Open Source Data Framework for LLMs
- Andrew Feldman: The Rise of Custom Foundation Models
If you enjoyed this episode, please support our work by encouraging your friends and colleagues to subscribe to our newsletter:
