Goal
Accurately extract key data from resumes
Challenge
Recruitment is a $200 billion industry with millions of job seekers uploading resumes and applying to jobs every day.
Resumes are growing increasingly more creative in style and presentation, making automating data extraction more difficult due to differing formats, OCR-based solutions which rely on template-based forms are limited.
The client is currently using an internal solution for resume extraction and data matching; however, their extraction model only achieves 90% accuracy and can only process text-based documents (.docx), not image-based ones (pdfs)
Outcome
Trained model with ~98% accuracy to extract key data from resumes regardless of format (image and text)