Building a Modern Search Engine: The Road Ahead
summary
Roadmap for 2026 commits OpenSearch to attract data scientists and AI engineers by compressing complexity so hybrid search and rerankers work out of the box. The repo and ML plugin show legacy bloat — a single plugin is a 380MB plugin — so Launchpad is shipping to hide the five API calls required to enable vectors. Priorities include simpler vector APIs, pull‑based ingestion, built‑in AB testing and migrations, workload management across hybrid workloads, and improving OpenAPI coverage (currently ~OpenAPI spec 70%).
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:32Need for new OpenSearch story
- 6:45Modern search communities
- 14:21Defining modern search
- 17:11Simplifying search engines
- 25:42OpenSearch as database
- 27:46Custom search APIs problem
- 31:22Modernizing data processing
- 33:30Focus on actions not results
- 35:59Building complete streets
- 39:42Q&A
related talks
18mZero To Hero: Building a Developer-First Search Experience With OpenSearch
31mNext-Gen Search: How Uber and the OpenSearch Community Built a Cloud Native 3.0
24mReimagining Search : Scaling OpenSearch Across an Enterprise - Platform C...
45mTomorrow’s SEO: How AI is Reshaping Search
37mThe Search for Search
40mBuilding a search platform 101 #bbuzz
45mBuilding amazing searches with Search API
36mUnlocking Smarter Typeahead Search