LLM Workflow / HR Automation
Recruiter Copilot : AI Screening Assistant
A full-stack AI screening tool that parses CVs, maps them to job requirements, and ranks candidates : with transparent evidence for every score.
Overview
Recruiter Copilot helps HR teams manage job descriptions, upload and parse CVs, run AI screening with candidate ranking, and export results to PDF or Excel. Every score is backed by extracted evidence so reviewers can challenge the AI's assessment.
Problem
CV screening is inconsistent. Different reviewers catch different things, quick keyword matching misses context, and there's no audit trail for why one candidate was ranked above another.
Goal
Create a transparent screening assistant that extracts relevant experience, maps it to job criteria, and supports human hiring decisions instead of replacing them.
Architecture
- Job description management with must-have and nice-to-have criteria.
- CV upload and parsing pipeline.
- AI extraction layer returning structured evidence per candidate.
- Ranking and comparison dashboard for reviewers.
System Flow
Input
Recruiter creates or uploads a job description.
Process
Candidate CVs are uploaded and parsed.
AI Layer
AI extracts experience, skills, and relevant evidence.
Storage/API
Dashboard shows criteria match with source snippets.
Review
Recruiter reviews, adds notes, and decides next step.
Tech Stack
Key Features
- CV text extraction and section classification.
- Candidate evidence mapped to job criteria.
- Reviewer notes and status pipeline (review / shortlist / reject).
- Export results as PDF or Excel.
AI / ML Component
- LLM-based structured extraction from CV text with schema validation.
- Criteria matching with supporting evidence snippets.
- Bias awareness reminders built into the review interface.
Data Flow
- 1Recruiter creates or uploads a job description.
- 2Candidate CVs are uploaded and parsed.
- 3AI extracts experience, skills, and relevant evidence.
- 4Dashboard shows criteria match with source snippets.
- 5Recruiter reviews, adds notes, and decides next step.
Challenges
- Handling diverse CV formats and layouts.
- Separating evidence extraction from final hiring decisions.
- Avoiding opaque AI scoring that reviewers can't challenge.
Solution / Trade-off
- Use AI as a review assistant, not an automated gatekeeper.
- Show evidence snippets so reviewers can verify or reject AI assessment.
- Keep score labels configurable per role or organization.
Result
Working full-stack application with CV upload, AI screening, ranking, and export. Ready for pilot with real job postings.
Screenshot / Demo Placeholder
/images/recruiter-copilot-placeholder.png
Replace this area with real screenshots, dashboard captures, architecture diagrams, or a short demo video once the asset is ready.
GitHub / Live Link Placeholder
What I Would Improve
- Add anonymized review mode to reduce unconscious bias.
- Add rubric versioning for compliance tracking.
- Add integration with applicant tracking systems.