OpenSearch Project Dublin - 3rd meetup

Hi Everyone,
mark your calendars for the
3rd OpenSearch Project Dublin Meetup - Friday January 17th 2025

Agenda
6.00pm Start (Drinks and snacks will be provided)
6.15 pm Lucas Jeanniot
Machine Learning Engineer at Eliatra
#The Myth of Unstructured Data: Leveraging Semantics to Power LLM-Search Engines."
In this talk, we’ll explore The Myth of Unstructured Data, delving into the fundamental differences between structured and unstructured data and their respective roles in powering LLM-driven search engines. While structured data is neatly organised into predefined formats, unstructured data—like images, free text, and audio—holds immense potential when analysed through the lens of semantics. By uncovering links and meaning hidden within unstructured content, we can transform it into rich, interconnected datasets that rival the utility of structured data.
This process not only builds powerful databases but also enhances LLMs' ability to deliver precise, context-aware search results, revolutionising RAG search systems.

7.00pm Break (Drinks and Snacks)

7.20 Fernando Rejon Barrera
CTO of Zeta Alpha
# So, You Want Vector Search on a Billion-Scale Collection?
Scaling vector search to a billion-scale collection is a balancing act of latency, cost, and quality. In this talk, I’ll share practical lessons from building a system at this scale, including insights into model selection, hardware requirements, and embedding strategies. I’ll also discuss tricks to address specific OpenSearch bugs encountered at this scale.

8.00pm Networking

8.30 Close