People and post matching method and device, equipment, storage medium and program product

Abstract

The invention discloses a person and post matching method and device, equipment, a storage medium and a program product, and relates to the technical field of human resource management, and the method comprises the steps: obtaining a post specification inputted by a user, and carrying out the analysis and standardization processing of the post specification, and obtaining a processed post specification; matching in a preset candidate database by combining the large model subjected to fine tuning training with a first requirement in the processed post specification to obtain candidate members; and in combination with a second requirement in the post specification after processing, scoring the candidate persons through the large model after fine tuning training to obtain final matching persons. According to the method, personnel are screened according to the first requirement in the processed post specification based on the large model subjected to fine tuning training, and then scoring and sorting are performed according to the second requirement for the screening result, so that more accurate post personnel recommendation is realized, and the accuracy, real-time performance and interpretability of personnel and post matching are improved.

Description

Technical Field

[0001] The present application relates to the field of human resource management technology, and in particular to a person-job matching method, apparatus, device, storage medium, and program product. Background Art

[0002] Current job-person matching technology mainly relies on traditional methods based on rules and keyword matching. It extracts keywords from resume text and job descriptions and calculates similarities to achieve preliminary screening of candidates. However, this type of method has many limitations. First, it relies on a preset keyword library or fixed rule model, which makes it difficult to adapt to diverse job requirements and talent expressions, and its generalization ability is poor. Second, traditional methods lack the ability to understand text semantics and contextual information, and cannot effectively explore the implicit ability characteristics and potential adaptability of candidates, resulting in inaccurate matching results. In addition, the existing system has weak dynamic update capabilities, making it difficult to respond to changes in job requirements in a timely manner, and there is a lag in matching results. At the same time, due to the lack of interpretability of the algorithm, companies lack trust in its decision-making process in actual applications, which affects the implementation and promotion of the job-person matching system. Summary of the Invention

[0003] The main purpose of this application is to provide a person-job matching method, device, equipment, storage medium and program product, aiming to solve the technical problem of how to improve the accuracy, real-time performance and interpretability of person-job matching.

[0004] To achieve the above objectives, this application proposes a person-job matching method, which includes:

[0005] Obtaining a job description input by a user, and parsing and standardizing the job description to obtain a processed job description;

[0006] By fine-tuning the trained large model and combining it with the first requirement of the processed job description, matching is performed in a preset candidate database to obtain candidate candidates;

[0007] Combined with the second requirement in the processed job description, the candidate is scored using the fine-tuned trained large model to obtain the final matching personnel.

[0008] In one embodiment, the step of parsing and standardizing the job description to obtain a processed job description includes:

[0009] Identify the type of the job description, and perform corresponding parsing and format conversion according to the type of the job description to obtain a standardized job description;

[0010] The standardized job description is parsed by fine-tuning the trained large model to obtain a processed job description.

[0011] In one embodiment, the step of parsing the standardized job description using the fine-tuned trained large model to obtain a processed job description includes:

[0012] Determining whether the input interface of the standardized job description is a preset job description recommendation interface;

[0013] If so, extracting standard coding data and weight information from the standardized job description as key parameters for person-job matching;

[0014] If not, a candidate job list is generated based on the job description content and the user's intention, and a preset database is searched based on the candidate job list to obtain job descriptions that meet the preset quality requirements;

[0015] If the query finds that the candidate job list data does not exist in the preset database, the steps are repeated: identifying the type of the job description, and performing corresponding parsing and format conversion on the job description to obtain a standardized job description.

[0016] In one embodiment, the step of combining the fine-tuned trained large model with the first requirement in the processed job description and performing matching in a preset candidate database to obtain candidate candidates includes:

[0017] Extracting necessary fields from the processed job description using the fine-tuned large model;

[0018] Determining a first requirement related to job matching based on the necessary fields, and converting the first requirement into a structured query language;

[0019] A query is performed in the preset candidate database according to the structured query language to obtain candidates that meet the first requirement.

[0020] In one embodiment, the step of combining the soft requirements in the processed job description and scoring the candidate using the fine-tuned trained large model to obtain the final matching personnel includes:

[0021] Based on the second requirement in the processed job description, define multi-dimensional labels for personnel matching, and define the scoring criteria and corresponding weights for each dimensional label;

[0022] Based on the scoring criteria of each dimension label and the corresponding weight, each candidate is scored by the fine-tuned large model to obtain the corresponding score of each candidate;

[0023] The final matching personnel are obtained by comprehensively considering the scores of each candidate.

[0024] In one embodiment, after the step of synthesizing the scores of the candidates to obtain the final matching person, the following steps are included:

[0025] Adjust the scoring criteria and weighting of each dimension label based on historical data;

[0026] Based on the adjusted scoring criteria and weights, the fine-tuned large model is optimized.

[0027] In addition, to achieve the above-mentioned purpose, the present application also proposes a person-job matching device, which includes:

[0028] A pre-processing module is used to obtain the job description input by the user, and parse and standardize the job description to obtain a processed job description;

[0029] A screening module is used to match the fine-tuned trained large model with the first requirement of the processed job description in a preset candidate database to obtain candidate candidates;

[0030] The scoring module is used to score the candidate personnel based on the second requirement in the processed job description through the fine-tuned trained large model to obtain the final matching personnel.

[0031] In addition, to achieve the above-mentioned purpose, the present application also proposes a person-job matching device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the person-job matching method as described above.

[0032] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium, and when the computer program is executed by the processor, the steps of the person-job matching method described above are implemented.

[0033] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the person-job matching method as described above.

[0034] This application proposes a person-job matching method, device, equipment, storage medium, and program product. The method includes: obtaining a job description input by a user, parsing and standardizing the job description to obtain a processed job description; using a fine-tuned trained large model in combination with the first requirement in the processed job description to match the job description in a preset candidate database to obtain candidate candidates; and using a fine-tuned trained large model in combination with the second requirement in the processed job description to score the candidate candidates to obtain the final matched candidates. This method is based on a fine-tuned trained large model, screening candidates according to the first requirement in the processed job description, and then scoring and sorting the screening results according to the second requirement, thereby achieving more accurate job personnel recommendations and improving the accuracy, real-time performance, and interpretability of person-job matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0037] Figure 1 A flowchart of the first embodiment of the applicant-job matching method is provided;

[0038] Figure 2 A flow chart of the second embodiment of the applicant-job matching method is provided;

[0039] Figure 3 This is a schematic diagram of the module structure of the person-job matching device according to an embodiment of the present application;

[0040] Figure 4 Schematic diagram of the device structure of the hardware operating environment involved in the person-job matching method in the embodiment of the present application.

[0041] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0042] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0043] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0044] The main solution of the embodiment of the present application is: obtain the job description input by the user, and parse and standardize the job description to obtain a processed job description; through the fine-tuning of the trained large model combined with the hard requirements in the processed job description, perform preliminary screening in the preset candidate database to obtain a preliminary matching candidate list; combined with the soft requirements in the processed job description, the fine-tuning of the trained large model is used to score the preliminary matching candidate list to obtain the final matching personnel.

[0045] In this embodiment, for ease of description, the following description is made with the person-job matching system as the execution entity.

[0046] Current job-person matching technology primarily relies on traditional rule-based keyword matching methods, which extract keywords from resumes and job descriptions and calculate similarities to achieve preliminary candidate screening. However, this approach has numerous limitations. First, it relies too heavily on pre-built models, resulting in poor generalization; its semantic understanding is limited, failing to tap into implicit skills; and its dynamic update capabilities are weak, resulting in delayed matching results and poor interpretability, leading to low enterprise trust.

[0047] In response to the unidirectionality, semantic limitations and lack of generalization of existing technologies, this application provides a two-way intelligent matching solution based on a large model. Based on the fine-tuned and trained large model, personnel are screened according to the hard requirements in the processed job description, and then the screening results are scored and ranked according to the soft requirements, thereby realizing the dual-mode collaboration of "finding people by job" and "finding people by people", and improving the dynamic utilization rate of the talent pool.

[0048] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or an electronic device capable of performing the above functions. The following uses a personal computer as an example to illustrate this embodiment and the following embodiments.

[0049] Based on this, the embodiment of the present application provides a person-job matching method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the applicant-job matching method.

[0050] In this embodiment, the person-job matching method includes steps S10 to S30:

[0051] Step S10: obtaining the job description input by the user, parsing and standardizing the job description to obtain a processed job description;

[0052] It is understandable that since the job descriptions written by different users have different formats and information expressions, there may be redundant or missing content. Therefore, executing step S10 can improve the standardization and consistency of job information and facilitate subsequent analysis and matching.

[0053] In this embodiment, when the user enters a job description (PDF or text), the system fine-tunes the trained large model to comply with the company's human resources data standards, parses and standardizes the job description, extracts key information, and prepares for subsequent matching.

[0054] Specifically, step S10 may include steps S11 to S12:

[0055] Step S11, identifying the type of the job description, and performing corresponding parsing and format conversion according to the type of the job description to obtain a standardized job description;

[0056] Specifically, this embodiment supports multiple modes of input for job descriptions. In addition to accepting job descriptions in plain text, the job matching system can also identify the type of job descriptions (such as PDF files or image formats) and call the corresponding parsing method for processing based on the type.

[0057] Specifically, the job description entered by the user is firstly identified by layout recognition, and the module contents such as text blocks, tables, figures, formulas, etc. are identified. Then, the contents of each identified module are parsed separately. For example, for PDF-type job descriptions, pdf-doclayout (a tool for identifying and analyzing the layout of PDF documents) is used to perform layout recognition and distinguish areas such as text blocks, tables, figures, formulas, etc.; for text block-type job descriptions, PaddleOCR (text recognition tool) technology is used for OCR (Optical Character Recognition) recognition to extract the text content in the image / scanned PDF as text; for table-type job descriptions, UniTable (unified table parser) technology is used for structured table recognition and content extraction. After completing the recognition and parsing of the content of each module, the information from different sources is uniformly converted into a standard text format while retaining the original structural information. Ultimately, all unstructured content is eventually converted into text with structural tags to facilitate subsequent semantic understanding and database storage.

[0058] Step S12: parsing the standardized job description through the fine-tuned trained large model to obtain a processed job description.

[0059] After structuring and standardization, the large language model (LLM) is used to semantically understand and enhance the job description content, improving its accuracy and intelligence for job matching.

[0060] Specifically, the job description standardized in step S11 and user instructions (such as "please generate a potential name for this position", "please extract implicit requirements", etc.) are input into the fine-tuned large model for parsing tasks.

[0061] Among them, the parsing tasks performed in the large model after fine-tuning training mainly include several aspects: First, check whether the input interface of the standardized job description is the preset job description recommendation interface.

[0062] If the input interface for the standardized job description is the preset job description recommendation interface, the response standard coding data and weight information of the job description are directly extracted as key reference information for person-job matching. The response standard coding data refers to the data that classifies and codes different positions according to industry standards.

[0063] If the input interface for the standardized job description is not the preset job description recommendation interface, the fine-tuned trained large model first extracts key entity information from the job description, such as the job title, department name, educational requirements, years of work experience, and skill keywords. Explicit conditions (directly stated requirements) and implicit conditions (requirements that require inference) are extracted from the original job description and user intent. For example, "bachelor's degree or above" and "familiarity with Python development" are explicit conditions, while "ability to withstand high-intensity work pressure" can be inferred to "adapt to fast-paced teamwork," and "responsible for project delivery" can be inferred to "possess project management capabilities," all of which are implicit conditions.

[0064] Then, based on the explicit and implicit conditions of the positions, potential position titles are generated, and finally a list of candidate positions is obtained.

[0065] Finally, through the list of candidate positions, query the high-quality job descriptions (including structured data established by standardized coding) stored in the preset database to extract key information and prepare for subsequent matching.

[0066] In addition, if no relevant candidate job list data is found in the preset database, the process returns to step S11 to obtain a standardized job description.

[0067] Through the above steps, data from different channels and in different formats are integrated to ensure the integrity and consistency of the data and provide rich information support for the model.

[0068] Step S20, fine-tuning the trained large model and combining it with the first requirement in the processed job description to perform matching in a preset candidate database to obtain candidate candidates;

[0069] It should be noted that the first requirement in this embodiment refers to the hard requirement, where the hard requirement refers to the conditions that must be met for the position, including but not limited to academic qualifications, majors, years of work experience, specific skills, certificates, etc.

[0070] It is understandable that due to the hard requirement of having clear judgment criteria and quantifiable characteristics, therefore, executing step S20 can generate SQL query statements by applying the fine-tuned trained large model given table information and description, and execute the query command for preliminary screening to quickly narrow the matching range.

[0071] Specifically, step S20 may include steps S21 to S23:

[0072] Step S21, extracting necessary fields from the processed job description using the fine-tuned trained large model;

[0073] Step S22: determining a first requirement related to the position matching based on the required fields, and converting the hard requirement into structured query language;

[0074] Step S23: query the preset candidate database according to the structured query language to obtain candidates that meet the first requirement.

[0075] In this example, the trained large model is fine-tuned to extract essential fields from the processed job descriptions. These essential fields include, but are not limited to, job title, age requirement, and educational requirements. These fields are then mapped to standard fields in SQL (Structured Query Language) queries using a standard adaptation module. These fields are then used as SQL query parameters to filter relevant data.

[0076] Specifically, the trained large model is first fine-tuned based on the extracted necessary fields to identify the hard requirements from the processed job descriptions, such as education, major, years of work experience, specific skills, and certifications. The extracted field values ​​are then normalized to align with the values ​​in the corresponding fields in the database. For example, "bachelor's degree or above" is normalized to "bachelor's degree," and "computer-related major" is normalized to "computer science."

[0077] Then, based on the hard requirements extracted from the job description, construct the corresponding SQL WHERE clause conditions. These clause conditions are designed to precisely match the candidate's education, professional background, years of work experience, specific skills, and certifications. For example, for education, you might use WHERE education_level >= 'Bachelor'; for major, you might use WHERE major LIKE '%Computer Science%'; for work experience, you might use WHERE years_of_experience >= 3 AND skill LIKE '%Python%'; for certifications, you might use WHERE certificates LIKE '%PMP%'.

[0078] Finally, the generated SQL statement containing clause conditions is used to query the preset candidate database and return candidate records that meet all hard requirements, thus completing the preliminary screening.

[0079] Through the above steps, the rigid requirements in the job description are accurately converted into SQL conditional statements, enabling the direct elimination of a large number of candidates who do not meet the rigid requirements and quickly narrowing the matching pool. For example, if a position requires a specific major and work experience, candidates who do not meet these requirements or who lack sufficient experience are immediately eliminated without further processing of these irrelevant records, significantly improving screening efficiency. Furthermore, through fine-tuning and training, the large model accurately grasps the semantics and logical relationships between the various requirements in the job description when understanding natural language and generating SQL. It can handle complex condition combinations, such as "having more than three years of Java development experience and familiarity with the Spring and Hibernate frameworks," ensuring that the generated SQL conditions accurately express these requirements, thereby ensuring accurate screening. Furthermore, the pre-set candidate database typically indexes frequently used for screening fields (such as education, work experience, and skills). When executing SQL queries based on these fields, the database can leverage the indexes to quickly locate records that meet the conditions, further improving screening speed. Furthermore, because the SQL statements generated by the large model filter based on precise fields and conditions, it leverages the advantages of database indexes to quickly narrow the pool while ensuring accurate results.

[0080] In step S30, the candidate is scored using the fine-tuned large model based on the second requirement in the processed job description to obtain the final matching personnel.

[0081] It should be noted that the second requirement in this embodiment refers to soft requirements, where soft requirements refer to personal qualities, abilities and traits that are difficult to quantify through specific data or hard indicators, including but not limited to knowledge, skills, quality and ability.

[0082] It is understandable that different positions have different soft ability requirements for candidates, and these abilities are difficult to measure directly through structured data. Therefore, executing step S30 can more accurately evaluate the comprehensive match between candidates and positions, thereby improving the intelligence level of talent matching and the quality of decision-making.

[0083] In this embodiment, each employee in the preliminary candidate list is scored. By scoring and ranking, the candidate's compatibility is comprehensively assessed, and the top three most compatible candidates (e.g., top 3) are recommended. Finally, the recommendation results are presented in a clear and intuitive manner, including the matching candidate's basic information, compatibility score, and specific matching factors, making it easier for companies to make decisions.

[0084] Specifically, step S30 may include steps S31 to S33:

[0085] Step S31: defining multi-dimensional labels for personnel matching based on the second requirement in the processed job description, and defining scoring criteria and corresponding weights for each dimensional label;

[0086] Step S32: Based on the scoring criteria and corresponding weights of the dimensional labels, the fine-tuned large model is used to score each candidate, thereby obtaining a score for each candidate.

[0087] Step S33: The scores of the candidates are comprehensively considered to obtain the final matching candidates.

[0088] Specifically, first, define the multi-dimensional labels for personnel matching and define the scoring weights of each dimensional label.

[0089] Currently defined dimension labels include, but are not limited to, academic qualifications, professionalism, stability, performance (e.g., performance results over the past three years), and age. Scoring criteria for each dimension can be set based on enterprise requirements. For example, the academic qualifications, performance, and age scoring criteria for each candidate's job description are directly based on business rules. For example, a doctoral degree is scored at 8-10 points, a master's degree at 6-8 points, a bachelor's degree at 4-6 points, and an associate degree or below at 1-4 points. When scoring the professionalism and stability of each candidate's job description, a fine-tuned and trained large model is used based on pre-set scoring rules and business logic to perform a comprehensive score. For example, the fine-tuned and trained large model is used to assess an employee's knowledge, skills, and experience in the relevant professional field. This model comprehensively considers factors such as professional course grades, professional certificates, and project experience.

[0090] Alternatively, the employee's stability can be judged based on his or her work experience and rational evaluation rate. For example, an employee who has worked in a company for more than 5 years and has no frequent job-hopping history can score 8-10 points, an employee who has worked for a company for 3-5 years and has changed jobs less frequently can score 6-8 points, an employee who has worked for a company for 1-3 years and has a certain number of job-hopping can score 4-6 points, and an employee who frequently changes jobs or has a shorter working life can score 1-4 points.

[0091] After setting the multi-dimensional scoring labels, the corresponding scoring criteria for each dimension, and the weighting, the fine-tuned trained large model is used to score each candidate in the candidate list, obtaining the corresponding score for each candidate. The top candidates with the highest scores are then selected as the final matches.

[0092] In this step, because the fine-tuned large model fully considers multi-dimensional labels and business rules in the scoring process, the comprehensive score can more comprehensively reflect the degree of match between employees and positions. The fine-tuned large model can comprehensively consider multiple dimensions such as education, professionalism, stability, performance, and age, avoiding the limitations of single-dimensional evaluation. In addition, different positions attach different importance to these dimensions. By setting reasonable weights, the fine-tuned large model can more accurately reflect the actual needs of the position, thereby improving the accuracy of the match. For example, for technical research and development positions, professionalism and education may be more important. By assigning higher weights to these two dimensions, the large model can screen out employees who better meet the requirements in these aspects.

[0093] Through the above-described method, a job description input by a user is obtained, parsed, and standardized to obtain a processed job description. A fine-tuned and trained large model is then combined with the first requirement of the processed job description to match candidates against a preset candidate database to obtain candidates. The fine-tuned and trained large model is then used to score the candidates based on the second requirement of the processed job description to obtain the final matched candidates. This method uses the fine-tuned and trained large model to screen candidates based on the first requirement of the processed job description, and then scores and ranks the screening results based on the second requirement, thereby achieving more accurate job recommendation and improving the accuracy, real-time performance, and interpretability of job matching.

[0094] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 2 After step S33, the person-job matching method further includes steps S331 to S332:

[0095] Step S331: adjusting the scoring criteria and the weighting of each dimension label according to historical data;

[0096] Step S332: Optimize the fine-tuned large model based on the adjusted scoring criteria and weights.

[0097] As the industry develops and job requirements change, the large model after fine-tuning training needs to adjust the scoring standards and weights of each dimension label at any time to ensure that it reflects the latest industry trends and changes in job requirements.

[0098] Specifically, relevant data is first collected from past successful and unsuccessful matching cases, including but not limited to standard job descriptions, key dimension weights, and recall query deadlines.

[0099] Then, through in-depth analysis of historical matching data, key factors influencing fit can be identified. Simultaneously, the scoring criteria for each dimension can be dynamically adjusted based on industry trends and evolving job requirements. For example, if the current market prioritizes innovation and problem-solving skills, the weighting of these abilities in the "Professional Ability" dimension should be increased.

[0100] In addition to adjusting the scoring criteria, the weighting of each dimension needs to be reassessed based on actual circumstances. For example, in certain high-tech industries, technical capabilities often take precedence over other soft requirements, so their weighting in the overall score should be appropriately increased.

[0101] After adjusting the scoring criteria and weights, the updated parameters can be applied to the optimization phase of the large model. Specifically, historical data that has not been used in training for the past six months is first prepared as evaluation data. A small amount of manually designed data is then used as training data (input includes "industry description," "job title," "job description," "job requirements," "historical scoring criteria and weights," "recall rate under historical scoring criteria and weights," "comprehensive description of recall cases," and "comprehensive description of mismatched cases"; the output is the new scoring criteria and dimension weights). This serves as seed data, and data augmentation is performed using the large model. Multiple base models are selected (based on parameter count and model architecture) and their performance and generalizability on the evaluation data are tested. The augmented data is then used to fine-tune the series of base models, evaluating their performance and generalizability on the evaluation data before and after fine-tuning, thereby verifying the validity of the training data. If model performance improves, the fine-tuned model is used to further augment the data, manually verifying the augmented data, and finally, using supervised fine-tuning, the model learns the latest evaluation system.

[0102] This example method continuously incorporates the latest historical data and analyzes successful and unsuccessful matching cases to gradually optimize the scoring system and model performance. As data accumulates and the model is iterated, the system's matching accuracy will continue to improve, ultimately achieving efficient and intelligent matching of employees and positions.

[0103] Next, the above embodiment method will be illustrated with a specific scenario:

[0104] Assume that a technology company is looking for an experienced Python engineer to join its development team. The company's human resources department uses the person-job matching system based on this embodiment to assist in talent screening and matching.

[0105] First, the system parses and standardizes the job description entered by the user to obtain the following structured information:

[0106] Job title: Python engineer;

[0107] Core responsibilities: backend service design and development, writing high-quality code, solving technical problems, and optimizing system performance;

[0108] First requirement (hard requirement): education (bachelor's degree or above in computer-related major), years of experience (more than 3 years), skills (Django / Flask);

[0109] The second requirement (soft requirement): communication skills, teamwork spirit, learning ability, and problem-solving ability.

[0110] The system then initially screened five candidates based on hard criteria: A, B, C, D, and E. The soft skills assessment phase then began. Using the candidate's resume and interview transcript as input, the model scored them based on the following dimensions (each with a maximum score of 10), as shown in Table 1:

[0111]

[0112]

[0113] Table 1

[0114] As shown in Table 1, the final scores are D>B>C>A>E, so the top three with the highest scores (D, B, C) are recommended as the best candidates.

[0115] If the system analysis later finds that successful employees generally score above 8 on "problem-solving ability"; that "teamwork," while important, is less impactful than expected; or that, in previously unsuccessful matches, most people scored low on "learning ability," the system will adjust the scoring criteria and weightings, as shown in Table 2:

[0116] Dimensions Original weight Adjusted weights Explanation of changes in scoring criteria Communication skills 25％ 20％ Lowering the weight may result in less impact than expected Teamwork 25％ 20％ Downgraded due to performance falling short of expectations Learning ability 20％ 30％ Significantly improved, reflecting the need for continuous learning Problem-solving skills 30％ 30％ Increase weight because it is highly correlated with performance

[0117] Table 2

[0118] Subsequently, the adjusted scoring criteria and weight configuration data were sorted out, resulting in two sets of data. One set was converted into training data format, while the other set served as a validation set to verify the accuracy and recall of the fine-tuned large model under the new scoring system. After confirming the model's stable performance, it was deployed to the online system for subsequent job matching tasks.

[0119] As illustrated in the above example, the company not only standardized job descriptions but also accurately assessed the soft skills of Python engineer candidates using a fine-tuned, trained large model. Furthermore, the system continuously learns from historical data, dynamically adjusting scoring rules and model parameters, thereby continuously improving the intelligence and accuracy of talent matching.

[0120] It should be noted that the above examples are only used to understand this application and do not constitute a limitation on the applicant's job matching method. More simple transformations based on this technical concept are all within the scope of protection of this application.

[0121] This application also provides a person-job matching device, please refer to Figure 3 , the person-job matching device includes:

[0122] The pre-processing module 10 is used to obtain the job description input by the user, and parse and standardize the job description to obtain a processed job description;

[0123] A screening module 20 is configured to match the fine-tuned trained large model with the first requirement of the processed job description in a preset candidate database to obtain candidate candidates;

[0124] The scoring module 30 is used to score the candidate personnel based on the second requirement in the processed job description through the fine-tuned trained large model to obtain the final matching personnel.

[0125] The person-job matching device provided in this application, utilizing the person-job matching method described in the aforementioned embodiments, can address the technical issues of improving the accuracy, real-time performance, and interpretability of person-job matching. Compared to the prior art, the beneficial effects of the person-job matching device provided in this application are the same as those of the person-job matching method described in the aforementioned embodiments. Other technical features of the person-job matching device are the same as those disclosed in the aforementioned embodiments and are not further elaborated here.

[0126] The present application provides a person-job matching device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the person-job matching method in the above-mentioned embodiment one.

[0127] Reference below Figure 4 , which shows a schematic diagram of the structure of a person-job matching device suitable for implementing the embodiments of the present application. The person-job matching device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), etc., as well as fixed terminals such as digital TVs and desktop computers. Figure 4 The person-job matching device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0128] like Figure 4As shown, the person-job matching device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory 1002 or programs loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the person-job matching device to operate. The processing device 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: an input device 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the person-job matching device to communicate with other devices wirelessly or wired to exchange data. Although the figure shows a person-job matching device with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems can be implemented or have alternatively.

[0129] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a read-only memory 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are performed.

[0130] The person-job matching device provided in this application, employing the person-job matching method described in the aforementioned embodiment, can address the technical issues of improving the accuracy, real-time performance, and interpretability of person-job matching. Compared to the prior art, the beneficial effects of the person-job matching device provided in this application are the same as those of the person-job matching method described in the aforementioned embodiment. Other technical features of the person-job matching device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0131] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0132] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0133] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, a computer program) stored thereon, wherein the computer-readable program instructions are used to execute the person-job matching method in the above-mentioned embodiment.

[0134] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0135] The computer-readable storage medium may be included in the person-job matching device; or it may exist independently without being assembled into the person-job matching device.

[0136] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the person-job matching device, the person-job matching device enables the following: to obtain the job description input by the user, and to parse and standardize the job description to obtain a processed job description; to match the first requirement in the processed job description with the fine-tuned trained large model in a preset candidate database to obtain candidate personnel; and to score the candidate personnel through the fine-tuned trained large model in combination with the second requirement in the processed job description to obtain the final matched personnel.

[0137] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0138] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0139] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0140] The computer-readable storage medium provided in this application stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned person-job matching method. This computer-readable storage medium can address the technical issues of improving the accuracy, real-time performance, and interpretability of person-job matching. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the person-job matching method provided in the aforementioned embodiments, and are not further elaborated here.

[0141] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned person-job matching method when executed by a processor.

[0142] The computer program product provided in this application can solve the technical problem of improving the accuracy, real-time performance, and interpretability of job matching. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the job matching method provided in the above embodiment, and will not be elaborated here.

[0143] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A person-job matching method, characterized in that: The person-job matching method includes: Obtaining a job description input by a user, and parsing and standardizing the job description to obtain a processed job description; By fine-tuning the trained large model and combining it with the first requirement of the processed job description, matching is performed in a preset candidate database to obtain candidate candidates; Combined with the second requirement in the processed job description, the candidate is scored using the fine-tuned trained large model to obtain the final matching personnel.

2. The person-job matching method according to claim 1, wherein: The step of parsing and standardizing the job description to obtain the processed job description includes: Identify the type of the job description, and perform corresponding parsing and format conversion on the job description to obtain a standardized job description; The standardized job description is parsed by fine-tuning the trained large model to obtain and store the processed job description.

3. The person-job matching method according to claim 2, wherein: The step of parsing the standardized job description by fine-tuning the trained large model to obtain a processed job description includes: Determining whether the input interface of the standardized job description is a preset job description recommendation interface; If so, extracting standard coding data and weight information from the standardized job description as key parameters for person-job matching; If not, a candidate job list is generated based on the job description content and the user's intention, and a preset database is searched based on the candidate job list to obtain the processed job description; If the query finds that the candidate job list data does not exist in the preset database, the steps are repeated: identifying the type of the job description, and performing corresponding parsing and format conversion on the job description to obtain a standardized job description.

4. The person-job matching method according to claim 1, wherein: The step of combining the fine-tuned trained large model with the first requirement in the processed job description and performing matching in a preset candidate database to obtain candidate candidates includes: Extracting necessary fields from the processed job description using the fine-tuned large model; Determining a first requirement related to job matching based on the necessary fields, and converting the first requirement into a structured query language; A query is performed in the preset candidate database according to the structured query language to obtain candidates that meet the first requirement.

5. The person-job matching method according to claim 1, wherein: The step of scoring the candidate by using the fine-tuned trained large model in combination with the second requirement in the processed job description to obtain the final matching personnel includes: Based on the second requirement in the processed job description, define multi-dimensional labels for personnel matching, and define the scoring criteria and corresponding weights for each dimensional label; Based on the scoring criteria of each dimension label and the corresponding weight, each candidate is scored by the fine-tuned large model to obtain the corresponding score of each candidate; The final matching personnel are obtained by comprehensively considering the scores of each candidate.

6. The person-job matching method according to claim 5, wherein: After the step of obtaining the final matching personnel by comprehensively considering the scores of the respective candidates, the following steps are included: Adjust the scoring criteria and weighting of each dimension label based on historical data; Based on the adjusted scoring criteria and weights, the fine-tuned large model is optimized.

7. A person-job matching device, characterized in that: The person-job matching device comprises: A pre-processing module is used to obtain the job description input by the user, and parse and standardize the job description to obtain a processed job description; A screening module is used to perform preliminary screening in a preset candidate database by fine-tuning the trained large model and combining it with the first requirement of the processed job description to obtain a list of preliminary matching candidates; The scoring module is used to combine the second requirement in the processed job description and score the list of preliminary matched candidates through the fine-tuned trained large model to obtain the final matched personnel.

8. A person-job matching device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the person-job matching method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the person-job matching method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the person-job matching method according to any one of claims 1 to 6 are implemented.

Images