JobPulse: An Advanced Personalized Job Recommendation System for IT Professionals

Abstract

In the dynamic landscape of the IT industry, the process of connecting employers with suitable candidates and vice versa has become akin to searching for a needle in a haystack. The abundance of information on job portals has created challenges for IT job applicants and employers alike. Job seekers often find it difficult to pinpoint opportunities that align with their skills and experiences, while employers struggle to identify candidates who possess the desired skill set and can seamlessly integrate into their company culture. The sheer volume of data sources compounds this issue, making it arduous to extract pertinent information tailored to the specific needs of both employers and job seekers. Traditional job-hunting processes often overlook crucial factors such as work experience and personality traits, further complicating the already intricate task of talent matching in the IT realm.

The system uses T5 models to extract skills, experiences and personality traits to match job seekers with suitable positions. The methodology is based on a multi-step strategy to handle the problem of job suggestion. This involves candidates filling out a questionnaire on their experiences, personalities, and qualifications. Three models are used to extract data on abilities, personality traits and experiences using pre trained T5 models. This approach enables precise and customized job recommendations. Initial results are promising, with match rates between job seekers and suitable positions significantly higher using our method. Our models, trained with pre-trained T5 models, achieving a significance accuracy. Precision, recall, and F1 score further demonstrate the model's performance. Combining neural networks with pre-trained language models enhances the speed and effectiveness of our recommendation system.