Memory efficient indexing to Elastic Cloud

Wednesday, November 20th, at 4:30 PM IST | Memory efficient indexing to Elastic Cloud using EmbedAnything built in Rust.

With the rise of vector databases, vector indexing is critical for powering LLMs, yet handling multivector embeddings in bulk can strain memory resources. Vector streaming addresses this by parallelizing the parsing, chunking, and embedding generation, and indexing data continuously, chunk by chunk, rather than all at once. This talk will mainly discuss how we can index embeddings to the elastic cloud using embed anything vector streaming adapters. What is vector streaming, and how does it improve performance? We will also share some benchmarks done with other similar platforms.

Speakers: Sonam Pankaj, GenerativeAI Evangelist Articul8,
Akshay Ballal, Msc. TU/e