kist-finder

To find KIST researchers by research topic offline, you query the markdown knowledge base with a natural-language search term. The system classifies the query as a topic and extracts matching candidates with evidence citations from local index files.

You can search KIST department members without network calls because the matching mechanism operates entirely offline. It uses grep and Read functions exclusively on markdown index files stored locally, ensuring deterministic behavior with no external dependencies.

The best way to identify a responsible leader for a technical domain is to submit a mixed or topic-based query. The system collects matching candidates and prioritizes responsible-level personnel, providing their department, role, and source URLs for verification.

The research directory search supports mixed queries combining names and departments. It classifies the query intent, aggregates candidates from the markdown indexes, and returns matched individuals alongside their departmental affiliations and evidence.

Offline markdown search returns deterministic results because it relies solely on grep and Read operations over static index files. Without network calls or external API execution, the matching logic processes the same local data consistently every time.

A limitation of searching the KIST keyword thesaurus with grep is that results are restricted to the pre-built markdown index content. If the index files are outdated or missing specific research terms, the search will not find those entries or require an index rebuild.