Overview
Explore the world of vector search engines in this 45-minute conference talk from Rust Linz March 2023. Delve into the capabilities of these emerging technologies, examine their inner workings, and discover why Rust is an ideal programming language for their development. Learn about neural search, similarity models, and vector search engine implementation from Arnaud Gourlay, a software engineer at Qdrant. Gain insights into concurrent programming challenges, distributed systems, and Raft consensus while understanding the evolution of search technologies and their practical applications.
Syllabus
Intro
Qdrant
Evolution of Search
Neural Search
Similarity Models
Vector Search Engine
Example
Additional Data
Implementation
Index
A N Search
Hnsw
payload filters
distance metrics
similarity function
naive product
CMD
Static vs Dynamic
CMD code
DOT benchmark
Architecture
Concurrent Programming
Mutex RWLock
Deadlocks
Deadlock Detection
Distribute
Raft Consensus
Conclusion
Taught by
Rust