An end-to-end pipeline for ingesting PDF documents, extracting atomic facts with verbatim evidence, and automatically reconciling claims across documents. Combines LLM extraction, deterministic rule engines, and grounded knowledge verification. Built for the SuperJoin on campus recruitment VIT Chennai.
nlp reconciliation nextjs gemini knowledge-graph fastapi pdf-processing fact-extraction llm superjoin
-
Updated
Sep 9, 2026 - Python