Vector Search and Databases at Scale
Unum Сloud
About speaker
Founder of Unum.cloud. Head of C++ Armenia 🇦🇲 community. Artificial Intelligence and Computer Science researcher. Ex-astrophysicist. Fluent in English, Russian & Armenian. For the past 15 years, he has been coding mainly in C++, CUDA, Python, and Assembly on x86/ARM.
About speakers's company
Unum is a deep-tech company designing the fastest data-processing software ever built. We work mostly in C++, CUDA and Assembly on x86 and ARM, our IO bypasses the Linux kernel, our DBMS doesn't need a file system, our Matrix Multiplications are fast for all Algebraic rings and sparsity levels, and our AI models are pre-trained on some of the biggest datasets ever assembled. Unum was established in 2015, is based in Yerevan, and works with a variety of companies worldwide.
4 July, 12:20, «Hall 3»
Abstracts
Vector Search databases appear on every corner. Most have already heard about Pinecone, Weaviate, Qdrant, and the Open-Source libraries that precede them ― Facebook's FAISS and Google's SCANN. We will look into the algorithms for approximate nearest neighbors search, profile them, and highlight the bottlenecks that stop most of them from scaling beyond a billion entries. This will help you navigate the increasingly complex space of vector and semantic search products, choose the optimal configuration parameters for indexing, the right neural network to produce embeddings, and the appropriate hardware for the task.
The talk was accepted to the conference program
