Jobsinc.ai: An Agentic AI Recruitment Platform Integrating Semantic Job Discovery, WhatsApp-Powered Hiring Workflows, and AI-Assisted Interview Evaluation

Recruitment remains a fragmented and inefficient process for both job seekers and hiring organisations. Candidates face platform siloing, keyword-based mismatch, and poorcommunication, while recruiters contend with disconnected tooling, manual CV processing, and inconsistent interview evaluation. This paper presents Jobsinc.ai, an end-to-end, multi-tenant agentic AI recruitment platform that addresses these failures across three integrated subsystems. The first is an applicant-centric job discovery pipeline combining NLP-based résumé parsing, multiplatform job aggregation, transformer-based semantic matching with cosine-similarity retrieval at a calibrated threshold τ= 0.42, and context-aware job recommendation. The second is a companycentric recruiter platform delivering AI-powered CV scoring and ranking, automated LinkedIn job posting, WhatsApp Business API-driven candidate communication via branded company channels, and a unified multi-tenant recruiter dashboard. The third is an AI-assisted interview evaluation subsystem comprising a Chrome browser extension for Google Meet session recording, AssemblyAI-powered transcription, and a two-stage LLM-based confidence-weighted checklist evaluation engine. Empirical testing across 87 test cases yielded a 93.9% pass rate, with all 8 measured operations meeting response-time targets. The formal confidenceweighted aggregate score S= siκi/ κi offers a stronger fairness guarantee than simple criterion averaging. Comparative analysis against LinkedIn, Indeed, and Rozee.pk confirms that no incumbent platform delivers the full combination of WhatsAppnative application, semantic CV matching, AI interview evaluation, and multi-tenant SaaS delivery in a single integrated system. Index Terms—résumé parsing, semantic job matching, cosine similarity, WhatsApp Business API, AI interview evaluation, multitenant SaaS, agentic AI, BullMQ, Pinecone, transformer models, recruitment automation