CareerBoost: A Hybrid RAG-NLP Job Recommendation Framework
Sanskruti Deshmukh, Aaryaman Bajaj · 2024
The proposed system presents a novel approach to job recommendation and resume enhancement, leveraging advanced Natural Language Processing (NLP) techniques and Retrieval-Augmented Generation (RAG). We introduce a comprehensive system that combines semantic search, FAISS-based similarity matching, and a fine-tuned GPT-2 model to provide personalized job recommendations and actionable resume improvement suggestions. The evaluation, conducted through offline metrics and online A/B testing with 5,000 users, demonstrates significant improvements over traditional keyword-matching approaches. The RAG-based system achieved a 25.8% increase in recommendation relevance (Precision@10), a 45.7% increase in user engagement, and a 41.2% increase in job application rates. Furthermore, users who applied the system’s resume enhancement suggestions saw a 9.0% improvement in subsequent job matches. These results highlight the potential of AI-driven systems to revolutionize online recruitment and career development, while also emphasizing the need for ongoing consideration of ethical implications and fairness in AI-assisted hiring processes.