Understanding Vespa with a Lucene mindset
summary
Vespa is a low latency platform that combines real time ingestion, search and analytics with built‑in ML and TensorFlow support and true partial updates so single fields can change without full reindexing. Architecture separates an application package and config service, stateless Java query containers, and a C++ content cluster for speed. Get started with Docker (~10 GB), Python 3, a schema.sd and services.xml, then run vespa deploy and use the feed client or HTTP documents. Query via YQL, use hits/order/filters and supply vector embeddings ingested externally for ANN search.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 2:22Understanding Vespa
- 5:14Vespa architecture
- 6:37Deployment and setup
- 8:56Data ingestion and updates
- 11:19Querying and interaction
- 14:19Vector search and conclusions
- 17:58Q&A and closing
related talks
40mVespa.ai’s Personalized Search: Advanced Ranking & Tensor framework
23mMachine Learning for Apache Solr, Elasticsearch & Vespa #bbuzz
22mTensor arithmetics in search and ranking for Ecommerce.
26mPersonalizing User Content - The Power of a Vector & Search DB
22mScaling ONNX and TensorFlow model evaluation in search
24mVector Search at Uber
28mFrom text search & recommendation to ads & online dating
01h 30mOptimizing Retrieval for Agentic Systems