For anyone reading this now: if you have the résumé as text, you can try a model that takes the field definitions as input instead of training one. I built ReDefiNER (HF - jungroo-ai/redefiner-0.8b · Hugging Face GitHub - Jungroo-AI/redefiner: A 0.8B model that extracts the entities you define in a plain-English rulebook. It runs on a CPU and is free under Apache 2.0. · GitHub , Apache 2.0, 0.8B, runs on a CPU). You list the types and rules in plain English:
[RULES]
Extract every entity of the supplied types. Return exact spans from the text.
[TYPE] DEGREE
[DESCRIPTION] A degree or qualification, such as BS, MS or PhD.
[TYPE] SCHOOL
[DESCRIPTION] A university, college or school.
[TYPE] COMPANY
[DESCRIPTION] An employer.
CO1 Do not mark schools or clients as COMPANY.
[TYPE] DATE_RANGE
[DESCRIPTION] A year or range of years, such as 2012-2016 or 2016 - Present.
It returns exact spans with a type and a confidence. Grouping them into Education and Experience entries is still your code, for example by section headings and line order. Spans stop at 16 tokens, so it suits fields, not whole paragraphs. You can also add precedence and examples for good accuracy on complex rules.