Timescale Newsletter ?? Create AI Embeddings in PostgreSQL
Hi friend,?
This week, we launched pgai Vectorizer, bringing embedding creation into PostgreSQL. With this PostgreSQL extension, you can treat embeddings as database indexes, streamline embedding synchronization, and experiment easily with embedding models—all within PostgreSQL, with one SQL command, no new tools required.
“Okay, but how do I use it?” you’re probably wondering. In this newsletter, we’ll show you how to simplify retrieval-augmented generation (RAG) system development with pgai Vectorizer and test different embedding models and chunking/formatting methods effortlessly for a smooth switchover when needed.
With pgai Vectorizer, you can turn PostgreSQL into your AI development platform of choice and fine-tune your embedding models’ performance without added complexity—helping you bring your AI application to market faster. Try it out—we’d love to hear your thoughts in our Discord community!?
And who doesn’t love a hackathon??????? We’re giving away $3,000 in prizes in our Open Source Challenge with pgai and Ollama. Scroll down?? for more details!
See you soon! ??
P.S. In case you missed it, our most-read article in the previous edition was Understanding Autoregressive Time-Series Modeling.
Hot Off the Press
FEATURED POST
Vector Databases Are the Wrong Abstraction Today’s vector databases disconnect embeddings from their source data. We should treat embeddings more like database indexes—here’s how.
READ
How to Automatically Create & Update Embeddings in PostgreSQL—With One SQL Query
Build RAG systems with ease and efficiency. Learn how to automate embedding creation and management with a single SQL query using pgai Vectorizer. Read more
READ
Which OpenAI Embedding Model Is Best for Your RAG App With Pgvector?
What if testing your data with a new OpenAI embedding model was as easy as running a SQL query? Learn more
READ
Which RAG Chunking and Formatting Strategy Is Best for Your App With Pgvector?
Pgai Vectorizer makes testing and experimenting with different RAG chunking and formatting approaches easy, helping you improve your AI system faster. Read more
领英推荐
WATCH
Auto Create and Sync Vector Embeddings in 1 Line of SQL
In this video, you’ll learn how to automatically create and sync vector embeddings using pgai Vectorizer. This PostgreSQL extension lets you create different versions of your embeddings using different models with a single SQL command and keeps them up to date for evaluation, A/B testing, and smooth switchover. Watch now
WATCH
What if Vector Embeddings Were Database Indexes?
Learn how to quickly create embeddings without leaving PostgreSQL using pgai Vectorizer in Timescale Cloud and a dataset from Sam Altman’s blog posts. Watch now
Join Us for the Open Source AI Challenge With Pgai and Ollama!
We ?? hackathons here at Timescale, so we’re really excited to announce the Open Source AI Challenge with pgai and Ollama in partnership with the DEV Community and Ollama! The challenge will run through November 10, and we’ll give away $3,000 in prizes. Your mandate is to build an AI application using PostgreSQL and two or more of these open-source extensions: pgvector, pgvectorscale, pgai, and pgai Vectorizer. Good luck!
Sharing the pgai Vectorizer ??
Tips From the Tiger Trenches
Sharing what our community reads, posts, or listens to!
Creating embeddings just got easier with pgai Vectorizer—no pipelines, no complex setups, just Postgres. Check out Uma's latest video for more details.
Timescale Careers
Team Timescale is growing! If you know someone who'd like to join our team—or learn more about life on Team Timescale ??—we're currently hiring (100 percent remote).?
?? Was this newsletter forwarded to you? Opt in to receive our newsletter and get these updates straight to your inbox!