Unstructured resumes, large candidate pools
4dot5 needed a recruitment platform that could parse resumes using NLP, match candidates to job openings semantically, track each candidate’s career progression state and produce screening reports as PDFs. Candidate pools were large, so the processing had to run asynchronously and at scale.
From parsed resume to ranked match
Daxtra NLP parses each resume and extracts structured data from it. An NLP resolver then matches candidates to jobs semantically, and AWS Lambda runs large-scale screening jobs asynchronously.
The application is Spring Boot on MongoDB, where 45+ collections give candidate and job documents a flexible data model. Quartz scheduling, persisted in MongoDB, drives the background workflows.
Around the matching
- Interview Mocha API integration for skills assessment
- AWS Chime for communication
- An alert framework for candidate and recruiter notifications
Resume screening on 4dot5’s platform is automated, from parsing through to semantic matching.