The Challenge
Finding one document in a large corpus is an indexing problem, not a browsing problem.
The Solution
A Next.js application over Elasticsearch: documents are indexed as they arrive, queries hit the index, and results come back ranked. Running live at particle.markbrown.space.
What I Built
Ranked Full-Text Search
Elasticsearch indexes the corpus and returns results by relevance, not filename.
Incremental Indexing
New documents are searchable as soon as they're added.
Outcome
- Relevant results in milliseconds across the corpus
Built With