MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641088169 A) filed by Karpaga Vinayaga College Of Engineering And Technology; Mr. J. Syed Raffi Ahamed; Dr. A. B. Hajira Be; and Ms. S. Charupriya on July 20, 2026, for An Intelligent Document Similarity And Ranking System For Automated Recruitment Matching Thereof.
Inventors include Mr. J. Syed Raffi Ahamed; Dr. A. B. Hajira Be; and Ms. S. Charupriya.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: The rapid influx of digital job applications has created a significant administrative bottleneck in corporate human resource pipelines, where conventional Applicant Tracking Systems (ATS) rely on rigid, keyword-based filtering that frequently overlooks highly qualified talent due to arbitrary terminology mismatches. To overcome these systemic limitations, the present invention discloses a data-driven, event-driven intelligent resume screening system and method thereof, designed to transition automated candidate evaluation from syntactic string matching to deep semantic intelligence. The platform establishes a secure digital infrastructure starting with an automated user authentication layer that initializes independent, multi-tenant database partitions to protect corporate data scopes. Recruiters interact with a centralized web application built on data visualization frameworks to dynamically create, configure, and update institutional job description matrices, establishing multi-variable target weights for specific technical competencies, domain experience, and adjacent capabilities. At its operational core, the platform integrates an asynchronous document ingestion pipeline that automatically captures uploaded candidate files, stripping formatting and extracting text layers into clean strings. Rather than executing linear timeline or word-frequency calculations, these extracted strings are systematically mapped alongside job configurations into a structured prompt array and processed via a Large Language Model framework—specifically optimized using the Gemini API—to execute a multi-step semantic reasoning chain. This intelligence engine evaluates professional proficiencies conceptually, recognizing synonym-based skill sets and framework adjacencies. To eliminate the opaque "black- box" nature of traditional automated screening, an integrated structured schema engine forces the model to serialize its qualitative evaluation into a machine-readable JSON object, mapping detailed text-based justifications, candidate strengths, and precise competency gap analyses directly alongside a calculated, weighted numeric fitness percentage score. This standardized payload is transmitted directly to a localized relational database engine, automatically updating candidate logs and schemas. The presentation tier subsequently polls this persistence layer to render real-time interactive graphical visualizers, candidate comparison matrices, and auditable reasoning frames across the recruiter interface, providing an objective, transparent, and legally defensible decision-support mechanism that significantly accelerates the recruitment lifecycle.
Disclaimer: Curated by HT Syndication.
