An intelligent agent system for college students' career planning based on large language model and multi-dimensional capability map

Abstract

This invention belongs to the field of artificial intelligence and career planning technology, and in particular, it is a college student career planning intelligent agent system based on a large language model and multi-dimensional ability graph. It includes a four-quadrant scenario interaction module, a two-way profile generation module, a graph-based career path planning module, and a four-dimensional weighted quantitative job-person matching module. This invention achieves intelligent and comprehensive interaction modes: based on a four-quadrant scenario design of "whether or not a resume is available, whether or not there is a job-seeking intention," an adaptive interaction mechanism is designed, embedding an AI structured query engine to proactively guide and extract information from students without resumes or who are career-challenged. It also achieves standardized and quantifiable profile construction: through a large language model, it achieves two-way construction of job ability requirement profiles and college student employment ability profiles, deconstructing and quantifying job and user abilities from multiple dimensions, while simultaneously completing the completeness and competitiveness score of the user's ability profile, allowing users to objectively understand their own abilities.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and career planning technology, specifically to a college student career planning intelligent agent system based on a large language model and a multi-dimensional ability graph. Background Technology

[0002] Career planning for university students is a core component of talent cultivation and graduate employment guidance in higher education institutions. High-quality career planning guidance can help students gain a clear understanding of themselves, accurately match career positions, and clarify their career development paths. Currently, the traditional career planning systems and guidance methods used by universities have many technical and application pain points in practical application and can no longer meet the career planning needs of university students in the new era.

[0003] The interaction mode is rigid and lacks proactive guidance. Existing systems often have a single input entry point, requiring students to provide a complete resume or a clear job search goal in order to use the system. For students who have no resume or are extremely confused about their career direction, the system lacks an effective information extraction and guidance mechanism and cannot provide effective planning guidance for such students. The assessment of students' self-awareness and positioning lacks objectivity. Existing methods rely heavily on subjective questionnaires to collect student information. Students find it difficult to deeply analyze their own interests and abilities through questionnaires, and the system is also unable to build accurate student ability profiles based on subjective information. Information asymmetry and one-sided understanding in the occupational field: the existing system lacks channels for conducting systematic and real job research, and cannot update core information such as job skill requirements and career development paths in a timely manner. Students have insufficient understanding of jobs in emerging fields and find it difficult to distinguish between real job requirements and gimmicky information. The current job matching technology is based on a single dimension and lacks dynamic adjustment. It is mostly a simple label comparison and does not conduct in-depth quantitative analysis from multiple dimensions such as basic requirements, professional skills, professional qualities and development potential. It also does not set dimension weights in combination with job characteristics. The matching results lack accuracy and cannot dynamically adjust the planning scheme according to the improvement of students' abilities. Career planning lacks practicality. Existing systems mostly only provide job recommendations without developing actionable plans to address the gap between students' abilities and job requirements, or constructing clear career development paths. As a result, students' career planning remains at the theoretical level and is difficult to implement. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a college student career planning intelligent agent system based on a large language model and multi-dimensional ability graph. It solves the problems of rigid interaction modes and lack of proactive guidance. Existing systems often have a single input entry point, requiring students to provide a complete resume or a clear job objective before using the system. For students without resumes or with extremely unclear career directions, the system lacks an effective information extraction and guidance mechanism, failing to provide effective planning guidance. Furthermore, student self-awareness and positioning assessment lack objectivity; existing methods rely heavily on subjective questionnaires to collect student information. Students find it difficult to deeply analyze their own interests and abilities through questionnaires, and the system cannot construct accurate student ability profiles based on subjective information. Finally, there is the issue of career information asymmetry and biased understanding; existing systems lack systematic real-world job research. The existing systems suffer from several problems. First, they fail to update core information such as job skill requirements and career development paths in a timely manner. Students lack sufficient understanding of emerging fields and struggle to distinguish between genuine job needs and hype. Second, job matching is often simplistic and lacks dynamic adjustment. Current matching technologies rely on simple tag comparisons, failing to conduct in-depth quantitative analysis across multiple dimensions, including basic requirements, professional skills, professional qualities, and development potential. They also lack weighting of dimensions based on job characteristics, resulting in inaccurate matching results and an inability to dynamically adjust planning schemes based on student skill development. Third, career planning lacks practicality. Existing systems often only provide job recommendations without developing actionable plans to address the gap between student and job skills or constructing clear career development paths. This leaves students' career planning at the theoretical level, hindering its implementation.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a college student career planning intelligent agent system based on a large language model and a multi-dimensional ability graph, including a four-quadrant scene interaction module, a two-way profile generation module, a graph-based career path planning module, and a four-dimensional weight-based quantitative person-job matching module. The modules work together to realize a closed loop of the entire process from user information extraction, two-way profile construction, multi-dimensional person-job matching to career planning scheme generation; the system can also be expanded to include a configuration report generation and dynamic adjustment module to improve the feasibility and dynamism of the planning scheme.

[0006] Four-quadrant scene interaction module This module serves as the system's information input entry point. Its core function is to dynamically switch information extraction logic and interaction modes based on four combined scenarios: whether the user has a resume, whether they have job-seeking intentions, etc. This breaks through the limitations of traditional systems with a single input and enables comprehensive guidance for students with different career awareness levels.

[0007] The module incorporates an AI-powered structured query engine. This engine is based on a large language model and has optimized prompt words. It has preset prompt words covering key dimensions such as personal information, educational background, interests and specialties, project experience, professional skills, and career aspirations. In scenarios where users do not have a resume, the engine guides users to answer questions in a proactive manner using natural language and automatically extracts effective information from the user's answers to complete the structured processing of information.

[0008] The specific interaction and information extraction logic for the four combined scenarios is as follows: With a resume and job preferences: The system directly calls the resume parsing engine to perform structured parsing of the resume uploaded by the user, extract core competency information, and perform multi-dimensional matching calculations with the user's specified preferred positions; For users with resumes but no job search intentions: the resume information is analyzed and a user competency profile is generated. The algorithm is then used to perform a global match in the job graph to recommend high-potential jobs that are a good fit for the user. No resume + job seeker intention: Activate the AI ​​structured inquiry engine to conduct targeted inquiries based on the user's desired job skills requirements, extract the user's core information, match it with the desired job, and generate a skills improvement action guide; No resume + no job intention: Activate the AI ​​structured inquiry engine to collect user information with open-ended and guided questions, extract core features, and then use algorithms to match suitable job directions, providing users with career exploration guidance from scratch.

[0009] Two-way portrait generation module This module is the core data support module of the system, realizing the bidirectional construction of job competency requirement profiles and college student employment competency profiles. It provides standardized and quantitative profile data for subsequent job-person matching, including job profile construction units and competency profile construction units: Job Profile Construction Unit: This unit uses web crawling technology to collect real recruitment data from various industries and positions, cleans, deduplicates, and standardizes the data; it then uses a large language model to extract and structure the information from the processed recruitment data, generating a standardized job competency profile that includes dimensions such as professional skills, certificate requirements, innovation ability, learning ability, and professional qualities, and quantifies and defines the indicators for each dimension. The Competency Profile Building Unit: Based on user information extracted from the four-quadrant scenario interaction module (resume parsing results or inquiry-based information), it uses a large language model to decompose and quantify the dimensions and features of user competencies, generating a college student employment competency profile that includes basic information, professional skills, project experience, interests and strengths, and development potential. At the same time, the module has a built-in scoring model that automatically scores the completeness of the competency profile (coverage of information in each dimension) and industry competitiveness (comparison of competencies with job seekers in the same industry / major), allowing users to clearly understand their own competency level.

[0010] Graph-based career path planning module This module is used to construct a job association graph and generate a visualized, multi-path career development plan for users based on the job-person matching results. It breaks through the limitations of traditional systems that only recommend single jobs, and includes a job graph construction unit and a career path generation unit. Job Map Construction Unit: Constructs two types of job relationship maps. The first is a vertical job promotion map, which outlines the job description, core competency requirements, and step-by-step promotion path (e.g., specialist-supervisor-manager-director) for each job, clarifying the competency improvement requirements at each promotion stage. The second is a horizontal job change path map, which analyzes the competency lineage between jobs (i.e., the general and transferable competencies required for different jobs) and configures at least two horizontal job change paths for a single job, covering job transition needs across fields and industries. Career path generation unit: Based on the matching results of the quantitative person-job matching module, combined with the user's career aspirations and ability profile, it extracts the corresponding vertical promotion path and horizontal job change path from the job map, and presents a clear career development path to the user through a visual interface; at the same time, it configures the ability requirements and improvement suggestions for each path node, so that the user can understand the development goals at each stage.

[0011] A quantitative job matching module based on four-dimensional weights This module is the core analysis module of the system, enabling multi-dimensional, quantitative, and precise matching of people and jobs, solving the technical problems of traditional matching methods that are limited in dimensions and lack weighting. The module sets four core matching dimensions: basic requirements, professional skills, professional qualities, and development potential. The specific evaluation content for each dimension is as follows: Basic requirements: These include hard requirements such as education level, major, graduating institution, and certificates; Professional skills: including practical skills for the job, such as professional operation skills, software usage skills, and business processing skills; Professional qualities include comprehensive qualities such as communication skills, teamwork skills, stress resistance, and a sense of responsibility; Development potential: This includes future development capabilities such as learning ability, innovation ability, adaptability, and career growth potential.

[0012] The module has a built-in job weight configuration model. Based on the characteristics and ability requirements of different jobs, it sets differentiated weight coefficients for four dimensions (e.g., technical jobs have a higher weight for professional skills, while management jobs have a higher weight for professional qualities). The system compares and analyzes the indicators of each dimension of the user's employability profile with the corresponding indicators of the job ability requirement profile, performs a weighted comprehensive calculation based on the weight coefficients, outputs a quantitative comprehensive person-job matching score, and completes the ability gap analysis of each dimension to identify the user's strengths and areas for improvement.

[0013] Report generation and dynamic adjustment module This module is the system's output module, enabling the structuring, implementation, and dynamism of career planning solutions. Its core function is to generate personalized career development reports based on job-person matching results and career development paths, and to support dynamic adjustments to the plans. The report includes: quantitative analysis of person-job matching, identification of core competency gaps, recommendations for suitable positions, a visualized career development path, short-term (1-2 years) / medium-term (3-5 years) career goals, phased implementation steps, and recommendations for capacity building resources. All content consists of specific, actionable, and implementable suggestions, addressing the problem of traditional planning schemes being superficial. Dynamic adjustment function: Preset quantitative assessment cycles and assessment indicators (such as quarterly / annual ability assessment, job intention assessment). Users can update their personal information in the system according to their own ability improvement and changes in career intention. The module will re-complete the person-job matching and career path planning based on the updated information to realize the dynamic iteration of career planning scheme. Report editing features: Intelligent polishing, integrity check, and manual modification functions are configured to meet users' personalized editing needs; it also supports one-click export to common formats such as PDF and Word, making it convenient for users to save, print, and use.

[0014] Compared with existing technologies, this invention provides a college student career planning intelligent agent system based on a large language model and a multi-dimensional ability graph, which has the following beneficial effects: 1. Intelligent and comprehensive interaction mode: Based on the four-quadrant scenario design of "whether there is a resume and whether there is a job intention", an adaptive interaction mechanism is designed and an AI structured query engine is embedded to achieve proactive guidance and information extraction for students without resumes and who are confused about their careers. This solves the problems of rigid interaction and narrow applicability of traditional systems and achieves comprehensive guidance for college students with different career awareness. 2. Standardized and quantified profile construction: Through a large language model, the profiles of job competency requirements and college students' employment competency are constructed in two directions. The job and user competency are decomposed and quantified in multiple dimensions. At the same time, the completeness and competitiveness of the user competency profile are scored, so that users can objectively understand their own abilities and solve the problem of one-sided and subjective self-awareness in traditional methods. 3. Precise and multi-dimensional matching of people and jobs: The matching of people and jobs is completed from four dimensions: basic requirements, professional skills, professional qualities and development potential. The dimensions are weighted according to the characteristics of the job and a weighted comprehensive score is performed. The quantitative matching results and ability gap analysis are output, which solves the problems of single dimension, no weight and low accuracy of traditional matching methods, and improves the scientificity and accuracy of the matching of people and jobs. 4. Visualized and multi-path career path: Construct a graph linking vertical promotion and horizontal job changes to provide users with a visualized career development path, covering job change needs across fields and industries. This breaks through the limitations of traditional systems that only recommend single jobs, allowing users to clearly understand their career development direction. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall modular architecture of the system of the present invention; Figure 2 This is a schematic diagram of the interaction logic of the four-quadrant scene interaction module of the present invention; Figure 3 This is a flowchart of the core implementation steps of the system of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example: Please refer to Figure 1-3 In this implementation plan, the core implementation steps of the system of the present invention are divided into four stages: construction of job profiles and correlation graphs, generation of user ability profiles in four-quadrant scenarios, four-dimensional weighted quantitative matching of people and jobs, and generation of personalized career development reports. Each stage is interconnected, realizing intelligentization of the entire career planning process. Phase 1: Job Profile and Relationship Graph Construction We crawl real recruitment data from all industries and positions across the entire internet, covering both traditional and emerging fields. We then clean, deduplicate, and standardize the data, removing invalid and false information. The finely tuned large language model is used to perform structured analysis on the processed recruitment data, extract core dimension information such as professional skills, certificate requirements, and professional qualities, generate standardized job competency profiles, and quantify the indicators of each dimension. Construct a job relationship graph, identify vertical promotion paths for each job, and form a vertical job promotion graph; analyze the competency relationships between jobs, configure at least two horizontal job transfer paths for each job, and form a horizontal job transfer path graph. Complete the overall construction of the job graph and enter it into the system database.

[0018] Phase Two: Generation of User Capability Profiles in Four Quadrant Scenarios The system receives a user's request and first identifies whether the user has uploaded a resume or filled in a job application, thus determining the user's four-quadrant scenario. For users with resumes, the system calls the resume parsing engine to perform structured parsing of the resume and extract core competency information; for users without resumes, the AI ​​structured query engine is activated to proactively query users through preset prompts, extract effective information from the answers and complete structured processing. The system utilizes a large language model to decompose and quantify the extracted user information to generate a profile of college students' employment abilities. The system's built-in scoring model is then used to complete the completeness score of the ability profile and the industry competitiveness score.

[0019] Phase Three: Four-Dimensional Weighting Quantification of Person-Job Matching Based on the user's job search intentions or algorithm recommendations, the system retrieves the competency profiles and weight configuration models for the target / recommended positions. From four dimensions—basic requirements, professional skills, professional qualities, and development potential—the user's employment ability profile is compared and analyzed with the job's ability requirement profile one by one to clarify the ability matching and gap in each dimension. By combining the dimensional weight coefficients of the job weight configuration model, the matching results of each dimension are weighted and comprehensively calculated to output a quantitative comprehensive person-job matching score, and generate a capability gap analysis report for each dimension.

[0020] Phase Four: Generation of Personalized Career Development Reports Based on the job matching results, the corresponding vertical promotion path and horizontal job change path are extracted from the job map, and a clear career development path is displayed to users through a visual interface. Based on the user's skill gaps and career goals, formulate short-term (1-2 years) and medium-term (3-5 years) career development goals, design phased and actionable implementation steps and skill enhancement suggestions, and generate a complete personalized career development report; The system allows users to configure evaluation cycles and quantitative evaluation indicators for reports, and supports users in updating information according to their own circumstances. The system then re-compiles the person-job matching and career planning based on the updated information, enabling dynamic adjustments to the plan. Users can also intelligently polish and manually modify the report, and export it to PDF / Word format with one click.

[0021] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A college student career planning intelligent agent system based on a large language model and multi-dimensional ability graph, characterized in that: The system includes a four-quadrant scenario interaction module, a two-way profile generation module, a graph-based career path planning module, and a four-dimensional weighted quantitative person-job matching module. The four-quadrant scenario interaction module dynamically switches information extraction logic and interaction modes based on four combined scenarios: whether the user has a resume and whether they have job-seeking intentions. The two-way profile generation module constructs structured job competency profiles based on real recruitment data and generates college student employment competency profiles based on user information, and completes the completeness and competitiveness scores of the competency profiles. The graph-based career path planning module constructs a graph linking vertical promotion and horizontal job changes, and generates a visualized career development path based on the person-job matching results. The four-dimensional weighted quantitative person-job matching module performs comparative analysis and weighted comprehensive scoring of person-job information from multiple dimensions, and outputs quantitative matching results.

2. The system according to claim 1, characterized in that: The job graph constructed by the graph-based career path planning module includes a vertical job promotion graph and a horizontal job change path graph. The vertical job promotion graph covers job descriptions and step-by-step promotion paths, while the horizontal job change path graph establishes connections between jobs through the ability lineage relationship, configuring at least two horizontal job change paths for a single job.

3. The system according to claim 1, characterized in that: The four-dimensional weighted quantitative job matching module comprises four dimensions: basic requirements, professional skills, professional qualities, and development potential. The module has a built-in weight configuration model for each job in the four dimensions. Based on the weight settings of the target / recommended job, the module performs weighted calculations on the user's abilities, outputs a quantitative comprehensive matching score, and completes the ability gap analysis in each dimension.

4. The system according to claim 1, characterized in that: The four-quadrant scene interaction module has four combined scenarios: having a resume and having job intentions, having a resume and not having job intentions, not having a resume and having job intentions, and not having a resume and not having job intentions. The module has an embedded AI structured query engine. In the scenario where the user does not have a resume, it guides the user to answer and extract effective information by actively asking questions through preset key dimension prompts covering personal information, educational background, interests and specialties, and project experience.

5. The system according to claim 1, characterized in that: The information extraction and matching logic of the four-quadrant scene interaction module for the four combined scenes is as follows: With a resume and job posting: The resume information is directly analyzed, and a four-dimensional matching calculation is performed with the desired job, outputting the matching degree and suggesting skills to be improved; With a resume but no job search intention: Analyze the resume information to generate a competency profile, and use algorithms to globally match it in the job graph to recommend highly suitable and potential jobs; No resume + job seeker: Extract user information through AI structured query engine, match with desired positions and generate targeted action guide; No resume + no job intention: The AI-powered structured query engine guides the extraction of the user's core characteristics, and the algorithm matches suitable job directions and provides introductory exploration guidance.

6. The system according to claim 1, characterized in that: The process of constructing the job competency profile of the bidirectional profile generation module is as follows: crawling and cleaning real recruitment data from various industries, extracting and structuring information through a large language model, and generating a standardized job profile that includes professional skills, certificate requirements, innovation ability, learning ability, and professional qualities; the college student employment ability profile is based on the results of user resume analysis or information extracted from inquiries, and completes dimensional decomposition and feature quantification through a large language model to achieve a completeness score and industry competitiveness score for the competency profile.

7. The system according to claim 1, characterized in that: The system also includes a report generation and dynamic adjustment module, which generates a personalized career development report based on job matching results and career development paths. This report includes job matching analysis, competency gap analysis, short-term / medium-term goals, and phased implementation steps. It also presets evaluation cycles and quantitative evaluation indicators to support dynamic adjustment of career planning schemes based on user competency improvement and changes in career intentions.

8. The system according to claim 7, characterized in that: The report generation and dynamic adjustment module is also equipped with report editing functions, including intelligent polishing, integrity checking, manual modification, and one-click export to PDF / Word format, to meet users' personalized editing and usage needs.

9. The system according to any one of claims 1-8, characterized in that, The system's large language model is a lightweight, finely tuned open-source large language model. It has undergone prompt word engineering optimization and industry data fine-tuning for the career planning scenario of college students, improving the accuracy of job information parsing, user information extraction, and job matching analysis.

Images