Aureon: Reconcile messy financial data with 99% accuracy

Hi everyone! 👋 I'm Pratik, the builder behind Aureon. Spending 1.5 years at FactSet, I saw that the global financial system isn't held together by APIs it's held together by "human glue." Armies of analysts manually reconcile messy spreadsheets because legacy software breaks the moment a data format changes. I built Aureon to fix this. It is a "Glass Box" AI engine that reads trade data like a human analyst but reconciles it at machine speed. Key features: Universal Ingestion: Drag-and-drop messy PDFs/CSVs. Hybrid Brain: Uses deterministic rules for speed and LLMs for complex reasoning (e.g., "T+3 settlement" notes). Audit Trail: Explains why a match was made. I'm currently building this solo and piloting with Indian custodians. I'd love to hear your feedback on the "fuzzy matching" logic!