HireX

HireX is an AI-Powered Transparent & Bias-Free Hiring System that transforms recruitment with cutting-edge technologies. It ensures fairness, accuracy, and transparency in candidate evaluation by combining Agentic AI, NLP, and bias mitigation into a scalable, real-time hiring pipeline.

Project Details

Designed an AI-driven JD-to-CV matcher using SBERT embeddings (786-D) to project candidates and job descriptions into semantic hyperspace, achieving 92% accurate cosine similarity-based matching.
Integrated IBM AIF360 into the pipeline for real-time bias detection and mitigation, ensuring fairness and compliance in candidate shortlisting.
Built a modular, multi-agent architecture via LangGraph & N8N that independently analyzes skills, projects, and experiences, validates claims against external sources, and aggregates scores transparently.
Implemented automated verification by cross-referencing candidates’ code repositories, portfolios, and social presence, reducing fraudulent claims and enhancing recruiter trust.
Developed an explainable AI Transparency Dashboard enabling recruiters to query and understand model decision-making at every step of the pipeline.
Optimized performance with Apache Kafka event-streaming and Python multithreading, enabling parallel processing of 100+ resumes daily with end-to-end latency as low as 22 seconds.
Engineered a high-speed, horizontally scalable backend (Flask + MongoDB) with load balancing, capable of processing 10,000+ resumes per day and future-proofed for 1M+ resumes/month.
Frontend recruiter dashboard (React + Node.js) allows intuitive interaction with AI insights, bias audit logs, and candidate rankings in real time.
Achieved 85% faster resume screening and 30% higher candidate-job matching accuracy compared to baseline systems.
Planned future enhancements include multi-language resume parsing for global reach and AI-powered fraud detection (deepfake detection for document and portfolio validation).