About the project
DigiTechBridge360 is an AI-driven recruitment marketplace connecting tech talent with employers through semantic "skill gap" analysis and automated upskilling recommendations.
My role: Sole developer, from the resume parsing pipeline to the AI orchestration engine.
DigiTechBridge360 is a dual-sided recruitment platform designed to solve the inefficiency of matching tech talent with employers. Unlike traditional platforms that rely on keyword matching, this project utilizes Gen AI to perform semantic analysis of "skill gaps," prioritizing potential and adaptability over static resume parsing.
The platform creates a value-driven ecosystem for two key users:
Talent Portal: Allows candidates to upload PDF resumes which are automatically parsed. Talents receive instant feedback on their fit for specific roles and actionable "micro-upskilling" plans to bridge identified skill gaps.
Employer Portal: Enables recruiters to post jobs and pre-select candidates based on an intelligent compatibility score and potential analysis, significantly reducing manual screening time.
AI Orchestration Engine: Built on a serverless architecture, the system utilizes Google’s Genkit and Gemini models to handle complex data parsing, gap analysis, and RAG-based learning recommendations while ensuring strict GDPR compliance.
Main features
- 📄 AI Resume Parsing: Automated extraction of skills and experience from PDF CVs.
- 🧠 Skill Gap Analysis: Semantic comparison between talent profiles and job descriptions to identify missing competencies
- 🎓 Upskilling Recommendations: RAG-based suggestions for micro-courses to bridge specific skill gaps.
- 🔐 Role-Based Access: Secure authentication and distinct interfaces for Talents and Employers.
- 📊 Compatibility Scoring: Intelligent scoring system (0-100) and generated interview questions based on candidate gaps
- ⚡ High Performance: Optimized Next.js frontend with Incremental Static Regeneration (ISR) for fast load times.
- 🛡️ Privacy by Design: GDPR-compliant data handling with automated "Right to be Forgotten" implementation.
Problems and solutions
Problem: Recruitment platforms match on keywords, so a candidate is filtered out for not writing a word rather than for not being able to do the job. Anyone who learned a skill under a different name, or who could learn it quickly, never reaches a human.
Solution: The platform reads resumes semantically and scores the gap between a candidate and a role rather than the overlap of their vocabulary. Potential and adaptability count, so a near fit surfaces instead of being silently dropped.
Problem: Being told you were rejected teaches a candidate nothing. They learn that they did not fit, but not what was missing or what would close the distance, so the next application repeats the same gap.
Solution: The analysis that scores the gap also names it. Candidates get the specific skills that separate them from the role and a micro upskilling plan to close them, which turns a rejection into a next step.
