Search, observability, and security software
Elastic
Elastic builds software for searching data, monitoring applications and infrastructure, and investigating cyber threats. Its Elasticsearch platform powers Search & AI, Elastic Observability, and Elastic Security, helping developers and practitioners retrieve information, support AI applications, and maintain operational systems. Customers can use cloud services, including hosted and serverless offerings, or manage the software themselves. Paid subscriptions use resource-based pricing, with some self-managed features available free.
Founded in 2012 by Steven Schuurman, Uri Boness, Simon Willnauer, and Shay Banon, Elastic grew around Elasticsearch, a distributed search engine built on Apache Lucene. Banon remains CTO; Ashutosh Kulkarni is CEO. Its engineering contributions include Better Binary Quantization, developed from insights in the RaBitQ research. The technique compresses stored vector dimensions into single bits while retaining higher precision for queries, reducing memory requirements for vector search.
Elastic reported $1.739 billion in revenue for its fiscal year ended April 2026, with subscriptions accounting for 94%. In August 2026, it completed its acquisition of Deductive AI to expand Elastic Observability’s production-incident investigation capabilities. Deductive AI’s agent gathers evidence and tests hypotheses across code, telemetry, and organizational knowledge to identify root causes. Existing customers remain supported while integration plans are developed.
3 talks
Newest first
2 speakers at AIE
Affiliations reflect their AIE appearances, not necessarily current employment.
Start here
- Information Retrieval from the Ground Up
Start here to understand the mechanics behind keyword, vector, and hybrid search before choosing a retrieval approach for RAG.
Philipp KrennAI Engineer World's Fair 2025
- Agentic Search for Context Engineering
Learn why fixed retrieval pipelines and narrowly scoped semantic search tools can fall short when agents need context.
Leonie MonigattiAI Engineer Europe 2026
- Vector Search Benchmark[eting]
Use this talk to assess vendor performance claims and design benchmarks around representative application workloads.
Philipp KrennAI Engineer World's Fair 2025
Messages from the stage
Retrieval beneath the RAG pipeline
Krenn's workshop uses Elasticsearch and Apache Lucene to explain how language analysis, tokenization, indexing, and scoring shape retrieval. It also demonstrates matching-fragment highlighting and semantic reranking.
Choosing tools and sources for agent context
Monigatti compares retrieval across databases, local files, agent skills, memory, and the web. Her workshop examines tool descriptions, shell interfaces, and context-window loading alongside the tradeoffs between database queries and file search.
Benchmark assumptions change the result
Krenn highlights read-only benchmark bias, counterintuitive HNSW filtering costs, and gradual performance regressions. Elasticsearch Rally and configurable tracks provide a practical route to reproducible testing.
Affiliations reflect each recorded session, not necessarily current employment.