scieee AI-readable full text Open interactive document viewer

Career Path Recommendation using Large Language Models

Iman Johary; Alexandru Mara; Tijl De Bie

Abstract

In the contemporary job market, efficient and precise resume analysis is essential for high-level tasks such as career path recommendations and job matching. This work introduces a novel approach to streamlining the process by harnessing the power of large language models. The primary objective is to develop a comprehensive pipeline for information extraction, data mapping, and subsequent utilization in downstream recommendation and matching tasks. Resumes are often parsed from different formats, resulting in unstructured and noisy text that poses a significant challenge for automated analysis. We aim to enhance the accuracy and efficacy of information extraction by leveraging recent developments in natural language processing, particularly large language models. The proposed pipeline begins with an information extraction module that efficiently retrieves key data points, including skills, qualifications, and experiences. Subsequently, a data mapping component maps different sections, such as job titles, places, and degrees, to a smaller taxonomy. The extracted information is then employed in downstream tasks, such as career path recommendations and job matching. For information extraction, we propose different methods, such as prompt engineering and LoRa. For data mapping, we experimented with fine-tuning and heuristic methods to increase accuracy. We evaluate the proposed pipeline using a rich dataset provided by VDAB, encompassing resumes in Dutch, French, and English. We further enhanced this dataset by annotating it to facilitate resume analysis and information extraction.

Full text

Career Path Recommendation using Large Language Models Iman Johary, Alexandru Mara, Jefrey Lijffijt, and Tijl De Bie Contact [email protected] This research was funded by the ERC under the EU's 7th Framework and H2020 Programs (ERC Grant Agreement no. 615517 and 963924), the Flemish Government (AI Research Program), and the FWO (project no. G0F9816N, 3G042220). Resume Personal information: Name: Iman Last Name: Johary Education: MSc Digital Electronic 2019-2021 Sharif University of Technology BSc Electrical Engineering 2015-2019 K. N. Toosi University of Technology Work Experience: Job title 2021-2022 Place: Company: Job Description: Internships: … Certificates: … Languages: … Skills: … Dataset Overview Original dataset: ➢Real-world data from VDAB. ➢Multilingual ➢Anonymized Annotated datasets: ➢Information Extraction 200 resumes ➢ESCO Mapping 1000 job experience ➢Motivation: Job market data is highly unstructured making it difficult to use in downstream tasks. ➢Problem: Transform unstructured resumes to structured and use in downstream tasks. Problem Overview “I worked at X company from 2015 to 2020. I worked in several data science projects, and I mostly used Python and Pytorch.” Work experience: Company: X Start year: 2015 End year: 2020 Title: Data Scientist Skills: Python, Pytorch { “education”: [ { “Degree”: “MSc Digital Electronics”, “University”: “Sharif University of Technology”, “Start Date”: “2019” “End Date”: “2021” } { “Degree”: “BSc Electrical Engineering”, “University”: “K. N. Toosi University of Technology” “start date”: “2015” “end date”: “2019”} ] } Prompts The following text is part of a resume, please restructure the Resume into JSON similar to the following format: Desired format. Instructions. Text LLM Evaluating using Information Extraction Dataset Model Title Acc Overall Acc Vicuna-13b 72.7 71.5 Vicuna-13b (Zero shot) 71.7 57.9 Llamma2-13b 72.2 53.0 Llama3.1-8b 78.9 78.7 Gemma2-9b 79.8 82.2 Information Extraction ESCO Mapping Downstream task ➢Improving performance: ➢Fine-tuning embedding model ➢Multilingual query ➢ESCO Alternative labels ➢Heuristic Ideas ➢Evaluating using ESCO Mapping Dataset Ablation Study Model ESCO code ESCO group Proposed method 41.5 58.3 W/o fine-tuning 39.9 56.5 W/o heuristic 35.4 53.4 W/o alt labels 41.2 57.8 W/o translation 32.8 52.2 Job Market Analysis ➢Impact of Career Breaks on Employment ➢Connection Between Education and the job market ➢Job Market Mobility ➢Additional considerations Job matching Model inputs Model outputs Job posting Parsed resume Job preference Career path recommendation