Resume screening method and related equipment

Abstract

The invention provides a resume screening method and related equipment, and the method comprises the steps: generating an information extraction instruction in response to description information of a first post; based on the first prompt information and the information extraction instruction, performing information extraction on the description information of the first post to obtain first content of the first post; and based on the first content, screening the resumes, and determining candidate resumes. In the information extraction process, information extraction prompting can be carried out through the first prompting information, the accuracy of the first content obtained through information extraction can be improved, the resumes are screened and candidate resumes are determined by using the first content obtained through information extraction subsequently, and the accuracy of the candidate resumes is improved. According to the resume screening method and device, the situation that due to manual subjective factor screening, the adaptability of the screened candidate resume and the first post is poor can be reduced, the adaptability of the screened candidate resume and the first post is improved, and therefore the resume screening effect can be improved through the screening process.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a resume screening method and related equipment. Background Art

[0002] With the development of the market in recent years, companies need to recruit people quickly, and HR also needs to respond to this short-term recruitment demand. The recruitment process mainly involves HR posting positions and collecting resumes from external recruitment channels; HR screening candidate resumes from a large pool of collected resumes, and HR interviewing qualified candidates; interviewers interview candidates based on their resumes, and after the interview, they evaluate and give feedback to the candidates to determine the interview results. Summary of the invention

[0003] The embodiment of the present application provides a resume screening method and related equipment, which is conducive to improving the resume screening effect.

[0004] In order to solve the above technical problems, this application is implemented as follows:

[0005] In a first aspect, an embodiment of the present application provides a resume screening method, comprising:

[0006] In response to the description information of the first position, generating an information extraction instruction;

[0007] Based on the first prompt information and the information extraction instruction, extract the description information of the first position to obtain the first content of the first position;

[0008] Based on the first content, the resumes are screened to determine the candidate resumes.

[0009] In a second aspect, the embodiment of the present application further provides a resume screening device, comprising:

[0010] A first generating module, used to generate an information extraction instruction in response to the description information of the first position;

[0011] An extraction module, configured to extract the description information of the first position based on the first prompt information and the information extraction instruction, so as to obtain the first content of the first position;

[0012] The screening module is used to screen the resumes based on the first content to determine the candidate resumes.

[0013] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the above resume screening method when executing the computer program.

[0014] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above resume screening method are implemented.

[0015] In a fifth aspect, an embodiment of the present application also provides a computer program product, including a computer program, which implements the steps in the above resume screening method when executed by a processor.

[0016] It can be seen that in this embodiment, there is no need for HR to manually screen candidate resumes from a large number of collected resume pools. Instead, after the recruiting unit / personnel releases the description information of the first position, an information extraction instruction is generated in response to the description information of the first position. The description information of the first position is extracted through the first prompt information and the information extraction instruction to obtain the first content of the first position. That is, in the information extraction process, the first prompt information can be used to perform information extraction prompts, which can improve the accuracy of the first content obtained by information extraction. The first content obtained by information extraction is used to screen resumes and determine candidate resumes to achieve candidate resume screening. This can reduce the situation where the screened candidate resumes have poor adaptability to the first position due to screening by manual subjective factors, and improve the adaptability of the screened candidate resumes to the first position. Therefore, through the above resume screening process, the resume screening effect can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0018] Figure 1 This is one of the flow charts of a resume screening method provided in an embodiment of the present application;

[0019] Figure 2 It is a schematic diagram of an interview solution provided in an embodiment of the present application;

[0020] Figure 3 This is a resume screening principle diagram in a resume screening method provided in an embodiment of the present application;

[0021] Figure 4 This is a schematic diagram of the principle of generating interview questions in a resume screening method provided in an embodiment of the present application;

[0022] Figure 5 This is an example diagram of an interview template provided in an embodiment of the present application;

[0023] Figure 6 It is a schematic diagram of a method for generating an interview summary evaluation provided by an embodiment of the present application;

[0024] Figure 7 It is a schematic diagram of an intelligent quality inspection provided by an embodiment of the present application;

[0025] Figure 8 It is a structural schematic diagram of a resume screening device provided in an embodiment of the present application;

[0026] Fig. 9 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] See also Figure 1 , Figure 1 1 is a flow chart of a resume screening method provided by an embodiment of the present application. The resume screening method of this embodiment can be applied to electronic devices, which can be mobile devices or non-mobile devices. Figure 1 As shown, the following steps are included:

[0029] Step 101: Generate an information extraction instruction in response to the description information of the first position.

[0030] A recruiting unit (recruiting enterprise) or a recruiter may publish the first position in a recruitment software (e.g., an artificial intelligence AI recruitment assistant, etc.) and publish the description information of the first position. The description information of the first position may include, but is not limited to, the ability requirement information for the first position, for example, including but not limited to the educational requirement information, educational background requirement information, professional ability requirement information, work experience requirement information, and general ability requirements (that is, basic ability requirements, for example, including but not limited to the ability requirements to use computers, the ability to use the software required for the first position, etc.) and at least one of the above. After the recruiting unit or the recruiter publishes the description information of the first position in the recruitment software, the method of the embodiment of the present application may generate an information extraction instruction in response to the published description information of the first position for information extraction. In addition, illustratively, the description information of the first position may also include but is not limited to the application time limit, work content, etc.

[0031] Step 102: Based on the first prompt information and the information extraction instruction, extract the description information of the first position to obtain the first content of the first position.

[0032] The first prompt information may be prompt information for guiding information extraction, and may also be referred to as information extraction prompt information. After the information extraction instruction is generated, information extraction may be performed in the description information of the first position based on the first prompt information in response to the information extraction instruction to obtain the extracted content, namely the first content.

[0033] As an example, at least one position capability requirement may be pre-set for the first position, which may also be called preset position capability requirement parameters, for example, including but not limited to at least one of age requirements, gender requirements, education background requirements, work experience requirements, patent capability requirements, and general capability requirements. As an example, the first content may include the content of the above-mentioned at least one position capability requirement extracted from the description information of the first position, which may represent the capability requirement information of the applicant in the published description information of the first position, for example, the professional requirement in the description information is computer science, and the work experience requirement is 2 years, and the corresponding information is extracted, and the professional capability requirement content obtained is computer science, and the work experience requirement content is 2 years.

[0034] As an example, the first prompt information can be information extraction prompt information of at least one position capability requirement. For each position capability requirement in at least one position, the corresponding first prompt information is obtained, that is, each position capability requirement has its own corresponding first prompt information. Based on the first prompt information of each position requirement, the description information of the first position is extracted to obtain the extraction result of each position capability requirement, that is, the content of each position capability requirement in the description information is obtained. In addition, it should be noted that prompt engineering: Prompt Engineering in English can refer to how to interact with the model without updating the model weights, and guide its behavior to obtain the desired results. In prompt engineering, the description of the task is directly input in the form of a question. Its typical working method is to convert one or more tasks into prompt-based inputs, and continuously guide the model to draw correct analysis conclusions. In other words, prompt engineering refers to stimulating the emergence of the model by designing special prompts. This method does not require additional training of the model, but only needs to guide the model to complete specific tasks by designing appropriate prompts. Prompt engineering is usually used to quickly solve new problems without updating model parameters. By inputting more reasonable prompt information, the model is guided to output more effective results. It is essentially a method to guide and stimulate model capabilities. In this embodiment, the first prompt information can be obtained, and based on the first prompt information, information extraction can be performed in the description information of the first position. For example, based on the first prompt information of at least one position capability, information extraction of at least one position capability requirement (which can be understood as an information extraction element) can be performed in the description information of the first position, that is, during the extraction process, the prompt information of the position capability requirement can be used to guide the information extraction of the position capability requirement, thereby improving the accuracy of information extraction.

[0035] Step 103: Based on the first content, the resumes are screened to determine the candidate resumes.

[0036] Applicants can upload their resumes in the recruitment software, so that the resumes of multiple applicants can be collected. For another example, the recruitment unit can publish a resume delivery mailbox on its website, and applicants can provide their resumes through the mailbox, so that the resumes of multiple applicants can be obtained.

[0037] After extracting the first content from the description information of the first position, the first content can be used to screen the resumes of multiple applicants to determine candidate resumes, thereby improving the compatibility of the screened candidate resumes with the first position.

[0038] It can be seen that in this embodiment, there is no need for HR to manually screen candidate resumes from a large number of collected resume pools. Instead, after the recruiting unit / personnel releases the description information of the first position, an information extraction instruction is generated in response to the description information of the first position. The description information of the first position is extracted through the first prompt information and the information extraction instruction to obtain the first content of the first position. That is, in the information extraction process, the first prompt information can be used to perform information extraction prompts, which can improve the accuracy of the first content obtained by information extraction. The first content obtained by information extraction is used to screen resumes and determine candidate resumes to achieve candidate resume screening. This can reduce the situation where the screened candidate resumes have poor adaptability to the first position due to screening by manual subjective factors, and improve the adaptability of the screened candidate resumes to the first position. Therefore, through the above resume screening process, the resume screening effect can be improved.

[0039] In some embodiments, based on the first prompt information and the information extraction instruction, extracting the description information of the first position to obtain the first content of the first position includes:

[0040] The description information, the first prompt information and the information extraction instruction are input into the large language model. Under the instruction of the information extraction instruction, the large language model extracts information from the description information according to the first prompt information to obtain the first content of the first position.

[0041] It should be understood that the information extraction instruction can be used to instruct the large language model (LLM) to extract information, so that the large language model can extract information from the description information according to the first prompt information, thereby obtaining the first content of the first position. In this embodiment, the large language model has different functions, and the information extraction function is an important function. By inputting the information extraction instruction into the large language model, the large language model can extract information under the instruction of the information extraction instruction. A large language model, also known as a large model, refers to an artificial intelligence model designed to understand and generate human language. The large language model is trained on a large amount of text data, and is characterized by its large scale, often containing billions of parameters, which are used to help it learn complex patterns in language data. In this way, in this embodiment, the first large language model is used to extract information under the instruction of the information extraction instruction to improve the effect of information extraction.

[0042] In some embodiments, based on the first content of the first position, the resumes are screened to determine the candidate resumes, including:

[0043] Parse the resume to obtain the second content in the resume;

[0044] determining a degree of match between the second content and the first content;

[0045] Based on the matching degree, the candidate resumes are identified from the resumes.

[0046] It can be understood that the second content is obtained by extracting information from the resume. The first content is the content extracted from the description information of the first position, which is the content of the recruiter for the first position, while the second content is the content extracted from the resume, which is the relevant content of the applicant. In the resume screening process, the resume is screened by determining the matching degree between the second content and the first content. The matching degree between the second content of the resume and the first content of the first position can indicate the matching degree between the resume and the requirements of the first position. By determining the candidate resume from the resume through the matching degree, the adaptability between the screened candidate resume and the first position can be improved.

[0047] In some embodiments, the second content of the resume may include the content of at least one position capability requirement extracted from the resume, which may represent the capability requirement information of the applicant in the resume, or may be understood as the specific content of at least one position capability requirement in the resume. For example, if the major in the resume is computer science and the work experience is 1 year, the corresponding information is extracted, and the content of the professional capability requirement obtained is computer science, and the content of the work experience requirement obtained is 2 years. The matching degree between the second content of the resumes of multiple applicants and the content of each position capability requirement may be determined. For example, the number of multiple applicants is N, and the number of at least one position capability requirement is M. For each applicant, M matching degrees may be determined (i.e., the matching degree between the content of each position capability requirement in the M position capability requirements in the second content and the content of the position capability requirement in the first content, i.e., one matching degree is the matching degree between the content corresponding to a position capability requirement in the second content and the content corresponding to the content in the first content), thereby obtaining N*M matching degrees.

[0048] After obtaining the matching degree between the second content and the first content of each applicant's resume among the multiple applicants, the matching degree is used to determine the candidate resume from the resumes of the multiple applicants. There are many ways to screen resumes using the matching degree, which are not specifically limited. For example, in one example, the resumes of n (positive integer) applicants with larger comprehensive matching degrees (for example, the average matching degree of M matching degrees corresponding to the applicants, the sum of matching degrees, etc.) among the resumes of the multiple applicants can be determined as candidate resumes. The comprehensive matching degree of the candidate resume refers to the comprehensive value of the matching degree between the second content and the first content of the candidate's resume. The comprehensive matching degree of the n candidate resumes is greater than or equal to the comprehensive matching degree of the resumes of other applicants, and the other applicants are applicants other than the n applicants among the multiple applicants.

[0049] In some embodiments, after determining the candidate resume, the method further includes:

[0050] Generate an interview template for the first position according to the first content of the first position, where the interview template includes multiple dimensional indicators;

[0051] Generate the second prompt information based on the candidate's resume and interview template;

[0052] Based on the second prompt information, a first interview question including multiple dimensional indicators is generated.

[0053] It should be understood that the above-mentioned dimensional indicators are indicators related to the first position, and can also be called interview assessment items. Each dimensional indicator can include corresponding sub-dimensional indicators, which can also be called interview assessment points or ability assessment points or knowledge assessment points, etc. For example, as an example, for the position of natural language processing (NLP) algorithm engineer, multiple dimensional indicators may include business problem indicators, basic problem indicators, job scenario simulation indicators, and programming question indicators. According to the first content of the first position, an interview template for the first position can be generated. Even if the first position is the same, different recruiters have different requirements, and the description information of the position is different, the first content obtained by information extraction may also be different. For different extraction results of the first position (i.e., different first contents), different interview templates can be generated, that is, personalized interview templates can be generated, so that the interview template is more suitable for the requirements of the recruiter for the first position.

[0054] After generating the interview template, the screened candidate resumes and the interview template can be used to generate the second prompt information. The second prompt information can also be called interview question generation prompt information, which is used to guide the generation of interview questions. The first interview question is generated based on the second prompt information. The first interview question may include interview questions corresponding to the above-mentioned multiple dimensional indicators. In addition, it should be noted that if multiple candidate resumes are screened, that is, the number of candidate resumes is multiple, then for each candidate resume in the multiple candidate resumes, the second prompt information can be designed in combination with the candidate resume and the interview template. The second prompt information can be interview question generation prompt information for multiple dimensional indicators in the interview template, and then the first interview question for the candidate is generated using the second prompt information. The generation of interview questions can be guided by the second prompt information to improve the quality of the generated interview questions.

[0055] In this embodiment, the second prompt information can be designed in combination with the candidate resume and the interview template. The second prompt information is used to guide the generation of interview questions to obtain the candidate's interview questions, thereby improving the quality of the generated interview questions.

[0056] In some embodiments, based on the second prompt information, a first interview question including multiple dimensional indicators is generated, including:

[0057] In response to the interview template, generating a test question generation instruction;

[0058] Inputting the second prompt information and the test question generation instruction into the large language model, and generating the candidate's initial interview question according to the second prompt information through the large language model under the instruction of the test question generation instruction;

[0059] Based on the third prompt information, the candidate's initial interview questions are screened to determine the candidate's first interview questions.

[0060] When generating an interview template, a question generation instruction can be generated to instruct the large language model to generate interview questions. After the second prompt information and the question generation instruction are input into the large language model, the large language model can generate initial interview questions for candidates based on the second prompt information under the instruction of the question generation instruction. It should be understood that the large language model has multiple functions, and the interview question generation function is one of its main functions. The large language model can generate interview questions under the instruction of the question generation instruction. In addition, the third prompt information (also referred to as the question screening prompt word) can also be obtained. The third prompt information can be used to guide the question screening. In this way, the initial interview questions for candidates generated by the large language model can be further screened under the guidance of the third instruction information, and the questions that do not meet the question screening requirements can be screened out to obtain the first interview questions, thereby improving the quality of the first interview questions.

[0061] In some embodiments, based on the third prompt information, screening the candidate's initial interview questions to determine the candidate's first interview question includes:

[0062] In response to the initial interview questions, generating test question screening instructions;

[0063] Inputting the third prompt information and the test question screening instruction into the large language model, and using the large language model to screen the candidate's initial interview questions according to the third prompt information under the instruction of the test question screening instruction, to obtain the candidate's second interview questions;

[0064] Perform quality inspection on the candidate's second interview question to obtain the quality score of the candidate's second interview question;

[0065] Determine the candidate's first interview questions based on the quality score of the candidate's second interview questions.

[0066] In this embodiment, in the process of screening the initial interview questions of the candidates based on the third prompt information, a question screening instruction can be generated in response to the initial interview questions of the candidates to instruct the question screening, and the third prompt information and the question screening instruction can be input into the large language model. The initial interview questions of the candidates are screened based on the third prompt information under the question screening instruction by the large language model. The third prompt information can guide the large language model to screen the questions and determine the second interview questions of the candidates to improve the screening accuracy.

[0067] In addition, the quality of the second interview questions of the screened candidates needs to be checked. It should be understood that the quality score of each question in the second interview questions can indicate the quality of the question. The higher the quality score, the higher the quality, and the lower the quality score, the lower the quality. In the process of using the third prompt information to screen the candidate first interview questions in this embodiment, the third prompt information is first used to screen the initial interview questions of the candidates to obtain the second interview questions of the candidates, and the quality of the second interview questions of the candidates is checked to obtain the quality scores of the second interview questions of the candidates. Based on the quality scores, the first interview questions of the candidates are determined to improve the quality of the obtained first interview questions of the candidates.

[0068] There are many ways to obtain the quality score of the candidate's second interview question, which are not specifically limited. Exemplarily, in the process of obtaining the quality score of the candidate's second interview question based on quality testing of the candidate's second interview question, you can first respond to the candidate's second interview question, generate a quality testing instruction, input the quality testing instruction into the large language model, and use the large language model to perform quality testing on the candidate's second interview question under the instruction of the quality testing instruction to obtain the quality score of the candidate's second interview question.

[0069] In some embodiments, determining the first interview question for the candidate based on the quality score of the second interview question for the candidate includes:

[0070] When the quality score of the candidate's second interview question is greater than the preset score, the candidate's second interview question is determined as the candidate's first interview question.

[0071] The preset score may be a preset score, which is not specifically limited. For example, it may be preset based on historical experience or actual needs. Since a larger quality score indicates a higher quality of the interview question, when the quality score of the candidate's second interview question is greater than the preset score, it indicates that the quality of the candidate's second interview question is high. In this case, the candidate's second interview question may be directly determined as the candidate's first interview question, thereby improving the quality of the candidate's first interview question.

[0072] In addition, exemplarily, when the quality score of the candidate's second interview question is less than or equal to the preset score, it indicates that the quality of the candidate's second interview question is low. In this case, the method may return to the step of generating the candidate's initial interview question according to the second prompt information through the large language model under the instruction of the question generation instruction to regenerate the initial interview question, update the initial interview question, and return to generate the question screening instruction in response to the initial interview question; input the third prompt information and the question screening instruction into the large language model, screen the candidate's initial interview question according to the third prompt information through the large language model under the instruction of the question screening instruction to obtain the candidate's second interview question, re-obtain the candidate's second interview question, update the candidate's second interview question, and return to the step of performing quality inspection on the candidate's second interview question to obtain the quality score of the candidate's second interview question to re-obtain the quality score and update the quality score, until the quality score of the candidate's interview question is greater than the preset score. At this time, the candidate's second interview question is determined as the candidate's first interview question to improve the quality of the determined first interview question.

[0073] In some embodiments, after determining the first interview question for the candidate, the method further includes:

[0074] Obtaining the interview record of the candidate during the interview process, wherein the interview record of the candidate includes the interview information associated with the first interview question of the candidate;

[0075] Based on the first content, the candidate's interview record and the qualifications for the first position, the interview result of the candidate is determined.

[0076] After determining the first interview question of the candidate, the interviewer may interview the candidate on the target interview question based on the first interview question of the candidate, and may obtain the interview record of the candidate during the interview process. For example, the interview information in the interview record may include the response to at least one interview question in the first interview question of the candidate (the interview question selected by the interviewer in the target interview question may be part of the test questions or all of the test questions in the target interview question), etc. The interview record may also include other information generated during the interview process, such as the candidate's consulting information (e.g., working hours, salary, etc.) and the interviewer's response to the candidate's consulting information, etc. Since the quality of the determined first interview question can be improved through the method of the embodiment of the present application, the first content extracted from the description information of the first position, the interview record associated with the candidate's first interview question, and the qualifications for the first position are subsequently used to determine the candidate's interview result and complete the interview with the candidate, so as to improve the interview effect.

[0077] In some embodiments, the method further comprises:

[0078] Monitor the candidate's interview process during the interview process;

[0079] When any of the K indicator parameters in the candidate interview process meets the corresponding preset conditions, a reminder message is output, where K is a positive integer.

[0080] That is, during the interview, the interview process can be monitored, for example, voice monitoring, video monitoring, etc. If an abnormality is detected in the interview, for example, if any of the K indicator parameters meets the corresponding preset conditions, indicating that the interview is abnormal, a reminder message can be output, and the reminder message can be used to remind the interviewer that the interview is abnormal, so that the interviewer can make corresponding interview adjustments based on the reminder message, so as to ensure that the interview goes smoothly.

[0081] It should be noted that each of the K indicator parameters has a corresponding preset condition. If any indicator parameter satisfies the corresponding preset condition, it means that the indicator parameter is abnormal, and a reminder message can be output. For example, different reminder messages can be output for different indicator parameters, that is, the reminder message can be the reminder message corresponding to the indicator parameter that meets the corresponding preset condition, so that the reminder message is more targeted and the reminder effect is improved.

[0082] In one example, the K indicator parameters may include but are not limited to at least one of an interview conversation indicator parameter (used to characterize the quality of the interview conversation), an interviewer indicator parameter (used to characterize the interviewer's behavior) and a candidate indicator parameter (used to characterize the candidate's behavior). Exemplarily, the interview conversation indicator parameter may include but is not limited to at least one of the conversation duration, the number of words in the conversation, the silence duration, and whether the interview is interrupted, reflecting the quality of the interview process, abnormal judgment: for example, for the conversation duration, if the conversation duration does not exceed the first duration threshold (the conversation duration meets the preset condition corresponding to the conversation duration), the silence duration exceeds the second duration threshold (the silence duration meets the preset condition corresponding to the conversation silence), or it is detected that the interview is no longer in progress (that is, the interview is interrupted, whether the interview is interrupted meets its corresponding preset condition), etc., then it is judged that an abnormality has occurred in the interview process. The interviewer indicator parameters may include but are not limited to at least one of the interviewer's emotions, interviewer's speaking speed, interviewer's response time, and whether the interviewer interrupts the interviewer. The interviewer's behavior is monitored and abnormal judgment is made: for example, if the interviewer is detected to have negative emotions, the speaking speed exceeds the preset speaking speed threshold, the number of interruptions exceeds the preset number threshold, or the response time of multiple answers exceeds the third preset time threshold, then the interview process is judged to be abnormal. The candidate indicator parameters may include but are not limited to at least one of the candidate's emotions and the candidate's speaking speed. Feedback on the candidate's interview experience and abnormal judgment are made: if the candidate is detected to have negative emotions, the speaking speed exceeds the preset speaking speed threshold, then the interview process is judged to be abnormal.

[0083] The following is a detailed description of the resume screening method using some specific embodiments. The scheme of the embodiment of the present application mainly includes four parts: resume screening (for example, automatic screening), interview question generation, interview summary evaluation, and intelligent quality inspection. The scheme of the embodiment of the present application can be applied to artificial intelligence (AI) recruitment assistants, such as Figure 2 As shown, it is a schematic diagram of the scheme of the embodiment of the present application. After the recruitment begins, HR releases the first position (i.e., job title), collects resumes, and the AI ​​recruitment assistant screens resumes. For example, based on the input description information of the first position (JD description), LLM is used to screen resumes from a massive resume pool for HR. In addition, the recommendation reason can also be given to improve the interpretability of the recommendation result. After the resume is evaluated as qualified, it enters the interview stage. HR arranges for an interview. After passing the interview, the interviewer will conduct an interview. During the interview, the AI ​​recruitment assistant can generate interview questions. For example, it can combine the description information of the first position, the establishment of candidates, interview templates, etc., and use LLM to automatically generate a set of interview questions with strong pertinence and high accuracy for the interviewer, assisting the interviewer to comprehensively examine the candidate. The AI ​​recruitment assistant can monitor the interview process and conduct intelligent quality inspection. For example, the detection operator is used to monitor the quality of the interview dialogue content, the attitude of the interviewer, the emotions of the candidate and other indicators in real time, to assist the recruiting unit in improving the quality of the interview and the interview experience of the candidate. After the interviewer completes the interview, the AI ​​recruitment assistant can conduct an interview summary and evaluation. For example, it can use a long text model combined with the first position's inspection points, qualifications, basic quality requirements, etc. to automatically generate the candidate's interview results (for example, it can include the results of passing or failing the interview, and can also include interview evaluations, etc.), and give interview feedback. If the candidate passes the interview, the next round of interviews will begin, for example, the next candidate will be interviewed. If the candidate fails the interview, the candidate's interview will end. After the candidate's interview is completed and no interview is scheduled, it can be determined whether the employing department has a recruitment demand. If not, and the position is full, the recruitment will end. If yes, and the position is not full, the next round of resume screening will be carried out.

[0084] The specific processing flow of each part is as follows:

[0085] (A), resume screening, the process is as follows Figure 3 shown.

[0086] The main steps of automatic screening of resumes are as follows: first, design a prompt project (corresponding to the first prompt information, i.e., the prompt words of the job ability requirements), the job ability requirements can be divided into four categories, and the definition of this category can be more accurately described by giving multiple examples of each ability requirement category, and the category definition and the description information of the first position form the first prompt information, and the first prompt information is input into the first large prediction model to extract the information of the job ability requirements, and the extraction result of the job ability requirements (corresponding to the first content) is obtained, and the candidate's resume (for example, a resume in PDF format, etc.) is parsed to obtain the parsed content of the job ability requirements (corresponding to the second content), for example, extracting educational background, work experience, etc., and finally, according to the extraction result and parsed content of the job ability requirements, the candidates are scored and sorted according to the job requirements, for example, the matching degree between the extraction result and the parsed content of the candidate's job ability requirements is determined (there are many ways to calculate the matching degree, which is not limited by the embodiment of the present application, and the matching degree of the embodiment of the present application can also be called similarity), the job matching scoring is completed, and the sorting is performed, and the scoring reason is given. According to the matching degree, the resume of the target candidate is determined in the resumes of multiple candidates.

[0087] Exemplarily, in the process of job matching and scoring, for each job ability requirement in at least one job ability requirement, the extraction result and parsed content of the job ability requirement are combined into a prompt word (prompt), so that a prompt word corresponding to at least one job ability requirement can be obtained. For example, the extraction result and parsed content of the education background requirement form a prompt word prompt1, the extraction result and parsed content of the work experience requirement form a prompt word prompt2, the extraction result and parsed content of the professional ability requirement form a prompt word prompt3, and the extraction result and parsed content of the general ability requirement form a prompt word prompt4. The prompt words corresponding to each job ability requirement are input into the large language model to obtain the matching degree (matching score) corresponding to each job ability requirement. For example, the education background matching score, work experience matching score, patent ability matching score and general ability matching score are obtained. The matching degree corresponding to each job ability requirement can be input into the seventh language model to obtain a comprehensive matching score and scoring reasons.

[0088] (B),Interview question generation process is as follows Figure 4 shown.

[0089] In the process of interview question generation, interview questions can be generated based on job competency requirements, candidate resumes, and template knowledge (interview template). Using a large language model to generate interview questions mainly includes the following steps:

[0090] First, we sort out the structured interview templates according to the job competency requirements, such as Figure 5The following is an interview template for a natural language processing (NLP) algorithm engineer position.

[0091] Then, we design prompt words based on the project description in the candidate’s resume and the examination points in the interview template, and input the prompt words into the large language model to generate the target interview questions:

[0092] For each inspection point under the interview template, a prompt word (corresponding to the second prompt information) is designed for the inspection point in combination with the project description in the candidate's resume, and a set of interview questions is generated using a large language model. The specific steps include: a. Generate a set of (for example, x) related interview questions (initial interview questions) based on the prompt word of the inspection point, b. Design a prompt word for question screening (corresponding to the third prompt information), and use a large language model to screen out the y interview questions that best meet the requirements from the above-generated set of interview questions, c. Design a quality scoring prompt word, and use a large language model to score the quality of this set (n) of interview questions, d. If the quality score (quality score) of the large language model is higher than the preset score (for example, 80 points, etc.), then output this set (n) of interview questions, that is, output the target interview questions, if the score is equal to or lower than the preset score, then repeat the above steps ac, and repeat the generation up to 3 times, and output the set (n) of interview questions with the highest score, that is, output the target interview questions.

[0093] (C), Interview summary evaluation, process as follows Figure 6 shown.

[0094] Interviews usually last a certain amount of time, and the number of words in the generated interview records may reach tens of thousands, which greatly exceeds the input length limit of the general LLM, so the LLM needs to be fine-tuned. LongLoRA technology can be used to train the long text large language model. The long text large model is used to summarize the interview process. First, the interview summary instructions are designed in combination with the extracted job competency requirements (or competency quality requirements) and the internal qualifications of the company, and the prompt words are combined with the interview record text to obtain the prompt words. Then, the prompt words are input into the long text large model to generate the interview evaluation, and the evaluation of the candidate's basic quality, knowledge and skills, and job matching is obtained.

[0095] (D), intelligent quality inspection, the process is as follows Figure 7 shown.

[0096] The interview process can be fully inspected (monitored), including but not limited to attitude inspection, error inspection, etc. If an abnormality is detected in the interview, the AI ​​recruitment assistant will send a reminder message to the interviewer. The indicator parameters for the inspection may include but are not limited to at least one of the following:

[0097] For the conversation content:

[0098] The main monitoring indicator parameters include: conversation duration, number of words in conversation, silence duration, whether the interview is interrupted, etc., to provide feedback on the quality of the conversation process.

[0099] Abnormal judgment: If the conversation duration does not exceed the threshold, the silence duration exceeds the threshold, or it is detected that the interview is no longer in progress, then it is judged that there is an abnormality in the interview process.

[0100] For the interviewer:

[0101] The main monitoring indicator parameters include emotions, speaking speed, response time, whether there is serious interruption, etc., to monitor the interviewer's behavior.

[0102] Abnormal judgment: If it is detected that the interviewer shows negative emotions, the speaking speed exceeds the threshold, the number of interruptions exceeds the threshold, the response time for multiple answers exceeds the threshold, etc., it is judged that there is an abnormality in the interview process.

[0103] For candidates:

[0104] The main monitoring indicator parameters include emotions, speaking speed, etc., and feedback on the candidate's interview experience.

[0105] Abnormal judgment: If it is detected that the candidate has negative emotions or the speaking speed exceeds the threshold, it is judged that there is an abnormality in the interview process.

[0106] In the solution of the embodiment of the present application, the AI ​​recruitment assistant can automatically score and sort resumes and give reasons for the scores, solving the problem that the current resume screening is easily affected by subjective factors and leads to low accuracy. The AI ​​recruitment assistant can ask targeted questions to candidates based on job competency requirements, candidate resumes, and template knowledge, and can examine candidate capabilities in a multi-dimensional, objective, and fair manner, solving the problem that it is difficult to unify the current standards for manual interviews; the AI ​​recruitment assistant can automatically evaluate the candidate's capabilities, saving the interviewer's time and improving recruitment efficiency; the AI ​​recruitment assistant can perform intelligent quality inspection on the interview process, which helps companies improve the candidate interview experience.

[0107] See also Figure 8 , Figure 8 800 is a structural diagram of a resume screening device 800 provided in an embodiment of the present application, which can implement the details of the resume screening method in the above embodiment and achieve the same effect. Figure 8 As shown, the resume screening device 800 includes:

[0108] The first generating module 801 is used to generate an information extraction instruction in response to the description information of the first position;

[0109] An extraction module 802 is used to extract the description information of the first position based on the first prompt information and the information extraction instruction to obtain the first content of the first position;

[0110] The screening module 803 is used to screen the resumes based on the first content to determine the candidate resumes.

[0111] In some embodiments, the extraction module 802 is specifically used to input the description information, the first prompt information and the information extraction instruction into the large language model, and the large language model extracts the description information according to the first prompt information under the instruction of the information extraction instruction to obtain the first content of the first position.

[0112] In some embodiments, the screening module 803 includes:

[0113] A parsing module, used for parsing the resume to obtain the second content in the resume;

[0114] A matching degree determination module, used to determine the matching degree between the second content and the first content;

[0115] The resume determination module is used to determine candidate resumes from the resumes based on matching degree.

[0116] In some embodiments, the apparatus further comprises:

[0117] An interview template generating module, used to generate an interview template for the first position according to the first content of the first position, wherein the interview template includes multiple dimensional indicators;

[0118] The first generation module is used to generate the second prompt information according to the candidate resume and the interview template;

[0119] The second generating module is used to generate a first interview question including multiple dimensional indicators based on the second prompt information.

[0120] In some embodiments, the second generation module includes:

[0121] A first generating unit, configured to generate a test question generating instruction in response to the interview template;

[0122] A second generating unit is used to input the second prompt information and the test question generating instruction into the large language model, and generate the initial interview question for the candidate according to the second prompt information through the large language model under the instruction of the test question generating instruction;

[0123] The screening unit is used to screen the initial interview questions of the candidate based on the third prompt information to determine the first interview questions of the candidate.

[0124] In some embodiments, the screening unit comprises:

[0125] A third generating unit, for generating a test question screening instruction in response to the initial interview question;

[0126] A second interview question determination unit is used to input the third prompt information and the test question screening instruction into the large language model, and screen the candidate's initial interview question according to the third prompt information through the large language model under the instruction of the test question screening instruction to obtain the candidate's second interview question;

[0127] A quality detection unit, used to perform quality detection on the second interview question of the candidate and obtain the quality score of the second interview question of the candidate;

[0128] The first interview question determination unit is used to determine the first interview question for the candidate according to the quality score of the second interview question of the candidate.

[0129] In some embodiments, the first interview question determination unit is specifically configured to:

[0130] When the quality score of the candidate's second interview question is greater than the preset score, the candidate's second interview question is determined as the candidate's first interview question.

[0131] In some embodiments, the apparatus further comprises:

[0132] A record acquisition module is used to acquire the interview record of the candidate during the interview process, wherein the interview record of the candidate includes the interview information associated with the first interview question of the candidate;

[0133] The interview result determination module is used to determine the interview result of the candidate based on the first content, the interview record of the candidate and the qualification of the first position.

[0134] In some embodiments, the apparatus further comprises:

[0135] A monitoring module is used to monitor the interview process of candidates during the interview process;

[0136] The output module is used to output reminder information when any indicator parameter among K indicator parameters in the candidate interview process meets the corresponding preset conditions, where K is a positive integer.

[0137] The resume screening device provided in the embodiment of the present application can perform each process of the resume screening method in the above embodiment. To avoid repetition, it will not be described here.

[0138] Fig. 9 A schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present application.

[0139] The electronic device 900 includes but is not limited to: a radio frequency unit 901, a network module 902, an audio output unit 903, an input unit 904, a sensor 905, a display unit 906, a user input unit 907, an interface unit 908, a memory 909, a processor 910, and a power supply 911. Those skilled in the art will appreciate that Fig. 9 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. In the embodiments of the present application, the electronic device includes but is not limited to a mobile phone, a tablet computer, a laptop computer, a PDA, a vehicle-mounted terminal, a wearable device, and a pedometer.

[0140] The processor 910 is configured to:

[0141] In response to the description information of the first position, generating an information extraction instruction;

[0142] Based on the first prompt information and the information extraction instruction, extract the description information of the first position to obtain the first content of the first position;

[0143] Based on the first content, the resumes are screened to determine the candidate resumes.

[0144] The processor of the electronic device 900 provided in the embodiment of the present application can implement each process of the resume screening method of the above-mentioned embodiments, and also has the same beneficial technical effects as the above-mentioned resume screening method embodiments, and the details will not be repeated here.

[0145] It should be understood that in the embodiment of the present application, the radio frequency unit 901 can be used for receiving and sending signals during information transmission or calls. Specifically, after receiving downlink data from the base station, it is sent to the processor 910 for processing; in addition, uplink data is sent to the base station. Generally, the radio frequency unit 901 includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc. In addition, the radio frequency unit 901 can also communicate with the network and other devices through a wireless communication system.

[0146] The electronic device provides users with wireless broadband Internet access through the network module 902, such as helping users to send and receive emails, browse web pages, and access streaming media.

[0147] The audio output unit 903 can convert the audio data received by the RF unit 901 or the network module 902 or stored in the memory 909 into an audio signal and output it as sound. Moreover, the audio output unit 903 can also provide audio output related to a specific function performed by the electronic device 900 (for example, a call signal reception sound, a message reception sound, etc.). The audio output unit 903 includes a speaker, a buzzer, a receiver, etc.

[0148] The input unit 904 is used to receive audio or video signals. The input unit 904 may include a graphics processor (GPU) 9041 and a microphone 9042, and the graphics processor 9041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The processed image frame can be displayed on the display unit 906. The image frame processed by the graphics processor 9041 can be stored in the memory 909 (or other storage medium) or sent via the radio frequency unit 901 or the network module 902. The microphone 9042 can receive sound and can process such sound into audio data. The processed audio data can be converted into a format output that can be sent to a mobile communication base station via the radio frequency unit 901 in the case of a telephone call mode.

[0149] The electronic device 900 also includes at least one sensor 905, such as a light sensor, a motion sensor, and other sensors. Specifically, the light sensor includes an ambient light sensor and a proximity sensor, wherein the ambient light sensor can adjust the brightness of the display panel 9061 according to the brightness of the ambient light, and the proximity sensor can turn off the display panel 9061 and / or the backlight when the electronic device 900 is moved to the ear. As a type of motion sensor, the accelerometer sensor can detect the magnitude of acceleration in each direction (generally three axes), and can detect the magnitude and direction of gravity when stationary, which can be used to identify the posture of the electronic device (such as horizontal and vertical screen switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), etc.; the sensor 905 can also include a fingerprint sensor, a pressure sensor, an iris sensor, a molecular sensor, a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, etc., which will not be repeated here.

[0150] The display unit 906 is used to display information input by the user or information provided to the user. The display unit 906 may include a display panel 9061, which may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), or the like.

[0151] The user input unit 907 can be used to receive input digital or character information, and to generate key signal input related to user settings and function control of the electronic device. Specifically, the user input unit 907 includes a touch panel 9071 and other input devices 9072. The touch panel 9071, also known as a touch screen, can collect the user's touch operation on or near it (such as the user's operation on the touch panel 9071 or near the touch panel 9071 using any suitable object or accessory such as a finger, stylus, etc.). The touch panel 9071 may include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the user's touch orientation, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into the contact point coordinates, and then sends it to the processor 910, receives the command sent by the processor 910 and executes it. In addition, the touch panel 9071 can be implemented using various types such as resistive, capacitive, infrared and surface acoustic waves. In addition to the touch panel 9071, the user input unit 907 may also include other input devices 9072. Specifically, other input devices 9072 may include but are not limited to a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which will not be described in detail here.

[0152] Furthermore, the touch panel 9071 may be overlaid on the display panel 9061. When the touch panel 9071 detects a touch operation on or near it, it transmits the information to the processor 910 to determine the type of the touch event. Then, the processor 910 provides a corresponding visual output on the display panel 9061 according to the type of the touch event. Fig. 9 In the figure, the touch panel 9071 and the display panel 9061 are used as two independent components to realize the input and output functions of the electronic device. However, in some embodiments, the touch panel 9071 and the display panel 9061 can be integrated to realize the input and output functions of the electronic device, which is not limited here.

[0153] The interface unit 908 is an interface for connecting an external device to the electronic device 900. For example, the external device may include a wired or wireless headset port, an external power supply (or battery charger) port, a wired or wireless data port, a memory card port, a port for connecting a device with an identification module, an audio input / output (I / O) port, a video I / O port, a headphone port, etc. The interface unit 908 may be used to receive input (e.g., data information, power, etc.) from an external device and transmit the received input to one or more elements within the electronic device 900 or may be used to transmit data between the electronic device 900 and an external device.

[0154] The memory 909 can be used to store software programs and various data. The memory 909 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory 909 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0155] The processor 910 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. It executes various functions of the electronic device and processes data by running or executing software programs and / or modules stored in the memory 909, and calling data stored in the memory 909, so as to monitor the electronic device as a whole. The processor 910 may include one or more processing units; preferably, the processor 910 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 910.

[0156] The electronic device 900 may also include a power supply 911 (such as a battery) for supplying power to each component. Preferably, the power supply 911 may be logically connected to the processor 910 through a power management system, thereby implementing functions such as charging, discharging, and power consumption management through the power management system.

[0157] In addition, the electronic device 900 includes some functional modules not shown, which will not be described in detail here.

[0158] Preferably, an embodiment of the present application also provides an electronic device, including a processor 910, a memory 909, and a computer program stored in the memory 909 and executable on the processor 910. When the computer program is executed by the processor 910, each process of the above-mentioned resume screening method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0159] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the resume screening method embodiment described above is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0160] The embodiment of the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the various processes of the above resume screening method embodiment and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0161] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0162] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0163] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. A resume screening method, characterized in that: include: In response to the description information of the first position, generating an information extraction instruction; Based on the first prompt information and the information extraction instruction, extract the description information of the first position to obtain the first content of the first position; Based on the first content, the resumes are screened to determine the candidate resumes.

2. The method according to claim 1, characterized in that The extracting information from the description information of the first position based on the first prompt information and the information extraction instruction to obtain the first content of the first position includes: The description information, the first prompt information and the information extraction instruction are input into a large language model, and the large language model, under the instruction of the information extraction instruction, extracts information from the description information according to the first prompt information to obtain the first content of the first position.

3. The method according to claim 1, characterized in that The screening of resumes based on the first content of the first position to determine candidate resumes includes: Parsing the resume to obtain the second content of the resume; determining a degree of match between the second content and the first content; Based on the matching degree, the candidate resume is determined from the resumes.

4. The method according to claim 1, characterized in that After determining the candidate resume, the method further includes: Generate an interview template for the first position according to the first content of the first position, wherein the interview template includes multiple dimensional indicators; Generate second prompt information according to the candidate resume and the interview template; Based on the second prompt information, a first interview question including the multiple dimensional indicators is generated.

5. The method according to claim 4, characterized in that The generating, based on the second prompt information, a first interview question including the multiple dimensional indicators includes: In response to the interview template, generating a test question generation instruction; Inputting the second prompt information and the test question generation instruction into a large language model, and generating an initial interview question for a candidate according to the second prompt information through the large language model under the instruction of the test question generation instruction; Based on the third prompt information, the initial interview questions for the candidate are screened to determine the first interview questions for the candidate.

6. The method according to claim 5, characterized in that The step of screening the initial interview questions for the candidate based on the third prompt information to determine the first interview question for the candidate includes: In response to the initial interview questions, generating a test question screening instruction; Inputting the third prompt information and the test question screening instruction into the large language model, and using the large language model to screen the candidate's initial interview questions according to the third prompt information under the instruction of the test question screening instruction to obtain the candidate's second interview questions; Performing a quality check on the second interview question of the candidate to obtain a quality score of the second interview question of the candidate; The first interview question for the candidate is determined according to the quality score of the second interview question for the candidate.

7. A resume screening device, characterized in that: include: A first generating module, used to generate an information extraction instruction in response to the description information of the first position; An extraction module, configured to extract the description information of the first position based on the first prompt information and the information extraction instruction, so as to obtain the first content of the first position; The screening module is used to screen the resumes based on the first content to determine the candidate resumes.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps in the resume screening method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the resume screening method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images