Article

Pizza, Vector Search, and more at the Data for AI Community Event Recap

•

November 21, 2024

•

7 min read

Summarize with ChatGPT

Last Tuesday evening we hosted the Data for AI After Party at our SF HQ. These are always great events filled with local community members looking to network and learn more about the latest news, tools, and tips for working with data and AI. The weather in the city was cold and blustery outside, but the pizza, drinks and conversations quickly warmed up as the presentations got underway.

We had a packed agenda with speakers from Elastic, Neo4j, Airbyte, DataStrato, and Roku. If you missed the event, you can check out the complete recording on our YouTube channel.

Susan sharing knowledge on search algorithms

Presentations kicked off with Susan Chang from Elastic and author of Machine Learning Interviews. I’ve seen Susan speak previously and she always does a great job of translating complex topics into relatable terms. For example, “raising the bar” and “going to the bar” are very different things. Keyword searching may think they are same, but in language models we are now able to differentiate them. Then, with vector searches, results in the same space such “high performing teams” could be returned as well to enrich the results.

Packed house

Next up, Nyah Macklin, from Neo4j explained how Graph technology is used at Google and how GraphRAG now offers improved accuracy for AI and enables similarity searches at the database level. This, GraphRAG, was a new concept for me and something I definitely want to dig into more. That's the great thing about these sort of community events; you learn so many things that inspire you keep exploring and learning more.

Nyah going deep on GraphRAG

After a quick slice of pizza - well maybe two; they were really good - Akriti Keswani from Airbyte took the floor to cover the need to unlock data from structured and unstructured data sources to fuel AI (After all, without data there is no AI) before giving a quick demo of developer options on Airbyte: from the no code AI Assistant to the low code Connector Builder designed to solve the long tail of connectors, and a brief overview of PyAirbyte, our python libraries for including data movement pipelines in your CI/CD and devops workflows.

Akriti talking about the AI challenge of unlocking data

After more pizza, Junping Du, CEO of Datastrato and Daniel Dai from Roku brought us home talking about unified data catalogs, fileset governance, and how Roku is leverage open source across their platform to improve performance.

Junping talking about file governance

Eventually, the evening drew to a close, wrapping up another great event. It’s so great to the see the vibrant ecosystem of community events return to San Francisco. There is a buzz back in the air around AI and the pace of innovation. If you haven’t joined one of the Data for AI or DataBytes events, sign up to be notified of the next. Despite how many slices of pizza I had, there’s plenty for everyone :)

‍

Limitless data movement with free Alpha and Beta connectors

Introducing: our Free Connector Program ->

The data movement infrastructure for the modern data teams.

Try a 30-day free trial