AURA: Autonomous Universal Recruitment Agent for Intelligent Large-Scale Job Application Automation

Online recruitment platforms have transformed the employment landscape, yet the burden on individual applicants remains disproportionately high. Applying to a single position through platforms such as LinkedIn typically requires three to five minutes of repetitive manual interaction, rendering large-scale job searching impractical. This paper presents AURA (Autonomous Universal Recruitment Agent), a Python-based autonomous recruitment framework that integrates Playwright browser automation, cookie-based session management, JSON-structured candidate profiling, and browser-native large language model interaction to execute end-to-end LinkedIn Easy Apply workflows without human intervention. In empirical deployment across the Artificial Intelligence, Machine Learning, and Data Science domains, AURA processed over 1,000 job postings across multiple geographic regions, achieving a 100% completion rate on all postings within the system's operational scope, at an average processing time of approximately five seconds per application—a 36-to-60-fold improvement over manual execution. Autonomous sessions sustained uninterrupted operation for three to four hours, exhausting all available regional job listings within a single session. A persistent JSON-based state ledger enables seamless execution resumption following interruptions. An integrated spaCy-based semantic filtering module performs resume-to-job-description alignment to improve application relevance. The framework further incorporates a headless browser interface for contextual LLM response generation to unseen screening questions, completing interactions within two seconds per query. AURA demonstrates that cost-effective, browser-native autonomous web agents can achieve production-grade recruitment automation, and raises substantive questions about the equitable governance of automated internet access.