AI Resume and Job Matching from Recruitment Group Chats
The challenge
A manpower agency received resumes and job orders as photos, PDFs and free-text messages in group chats. Staff retyped information into spreadsheets and matched candidates from memory, making search, review and follow-up slow and inconsistent.
What we delivered
- Intake from approved chat groups, files and images.
- OCR and structured extraction for candidate and job fields.
- Review states for identified, needs review and invalid records.
- Candidate-to-job matching with scores, reasons and condition checks.
- Placement tracking from recommendation to onboarding.
- Group-scoped permissions, audit history and intake analytics.
Integration and delivery
The workflow kept the original chat-based intake process and added a searchable operational layer beside it. Original files stayed with extracted fields so recruiters could verify and correct results instead of trusting an opaque AI response.
Technology
Next.js, React, TypeScript, FastAPI, PostgreSQL, Redis, OCR and role-based access control.
All candidate, employer and job data shown in the demonstration is fictional.