Multi-Format Candidate Data Parser and Normalizer

Design and implement a system that monitors multiple source directories containing candidate data in varying JSON schemas (and potentially XML or other formats). Each folder follows a consistent but distinct schema. The system should transform documents from all source formats into a single unified output format and store them in an output folder. Requirements: (1) Design a generic unified output format for candidate data (name, education, experience, skills). (2) Detect and process new, updated, and deleted documents in any input folder with minimal sync lag. (3) Support introducing a new folder with a different input schema without code changes. (4) Handle a large volume of documents efficiently. Example input schemas are provided with differing field names and structures for name, education (flat vs. nested school object), experience (flat vs. nested organization object), and skills (flat array vs. categorized objects).