Technical aspect test question generation method and device based on AI Agent and medium

Abstract

The embodiment of the invention discloses a technical aspect test question generation method and device based on an AI Agent and a medium. The method comprises the following steps: inputting post description and candidate resumes, and processing by using NLP and AI Agent to generate original technical tags; inferring a project technical architecture by using an AI Agent; performing technical label generation and expansion based on the original technical label and the project technical architecture to obtain a full-quantity technical label; and generating technical face test questions and interview feedback. The method has the following advantages: the technical label highly matched with the post is generated based on the NLP and the AI Agent, and the technical label is expanded, so that the accuracy and efficiency of the recruitment process can be remarkably improved; face test questions are generated according to the technical labels, and correlation and pertinence of the face test questions are ensured; the recruitment process can be dynamically fed back and optimized according to an interview result; the human resource management cost is remarkably reduced, and meanwhile the recruitment quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of natural language processing technology, and in particular to a method, device and storage medium for generating technical interview questions based on AI Agent. Background Art

[0002] Most current recruitment systems are based on simple keyword matching and rule engines to screen candidates and generate interview questions. Such systems mainly match skills and job requirements by parsing keywords in job descriptions (JDs) and candidate resumes. Typical existing technical solutions include:

[0003] Keyword matching system: This system relies on matching keywords in the job description and resume text. If the keywords in the resume match the key requirements in the job description, the candidate is considered to meet the basic requirements. The operating mechanism of this type of system is usually very simple, relying on string matching, predefined rules and fixed algorithms.

[0004] Interview question bank management system: Some recruitment systems come with a predefined interview question bank. Recruiters manually select interview questions from the question bank based on job requirements. These question banks are usually not automatically associated with job requirements, and the selection of questions depends on the recruiter's experience and judgment.

[0005] Manual resume screening and interview design: Although some systems provide resume parsing functions, the actual resume screening and interview design work is mostly done manually by human resources personnel or technical interviewers. The manual screening and design process is time-consuming and difficult to maintain consistency.

[0006] Although existing technical solutions can provide basic recruitment support functions in some scenarios, their limitations and shortcomings are also very obvious:

[0007] (1) Limitations of Keyword Matching

[0008] Unable to accurately identify semantics: Existing keyword matching systems only rely on specific words in resumes and job descriptions, and have difficulty understanding the true semantics behind these keywords. For example, some job descriptions may use certain terms or expressions, while the candidates in their resumes may use the same meaning but different expressions, which makes it difficult for the system to accurately match.

[0009] (2) Inability to distinguish importance: Existing systems do not have clear priorities or weights for matching keywords, and all keywords are considered equally important. This makes it difficult for the system to judge the importance and priority of skills when evaluating candidates, which may result in missing key skill matches or screening out unsuitable candidates.

[0010] (3) Insufficient interview question generation

[0011] Fixed question bank: Current interview question banks are usually pre-defined and cannot dynamically generate new interview questions. The content of interview questions is often out of touch with the real-time technical needs of the enterprise, lacking flexibility and pertinence.

[0012] Lack of personalization: Most interview questions are manually selected by recruiters based on job requirements, and it is impossible to generate personalized interview questions based on the candidate's technical background, work experience, and project experience. As a result, the interview questions for different candidates may be the same, and it is impossible to truly examine the candidate's fit with the specific position.

[0013] (3) Insufficient ability to infer technical architecture

[0014] Unable to infer the project technical architecture of the position: The existing recruitment system cannot infer the complete project technical architecture required for the position through the position description. Even if the keyword matching can identify the technical tools and programming languages ​​required for the position, the system cannot combine these technical tags to infer the architectural requirements of the position, making it difficult to generate more complete and in-depth technical interview questions.

[0015] Difficulty in expanding technical tags: After matching technical keywords in job descriptions, existing systems are unable to further expand relevant technical tags. They lack understanding of technical architecture and are unable to automatically generate expanded technical tags, which may lead to insufficient skill coverage.

[0016] (4) Inefficient recruitment process

[0017] Excessive reliance on manual labor: The existing system over-relies on the experience and judgment of recruiters in the process of resume screening and interview question generation, resulting in an inefficient overall process. Especially when recruiting for complex technical positions, the manual screening process is time-consuming and error-prone, which prolongs the recruitment cycle.

[0018] Lack of automation and intelligence: Although some systems can partially automate resume screening, they lack deep-level intelligent means, such as generating technical labels through AI analysis, inferring technical architecture, and automatically generating personalized interview questions, resulting in the overall process still relying heavily on manual operations.

[0019] (5) Weak feedback and optimization mechanisms

[0020] Difficulty optimizing the recruitment process based on interview results: Existing systems usually do not have an automatic feedback mechanism and cannot optimize the process based on feedback from interviewers or candidates. For example, when an interview question performs poorly, the system cannot automatically adjust or update the question bank, resulting in a long-term fixed question bank that is difficult to adapt to changes in job requirements. Summary of the invention

[0021] In view of the technical defects mentioned in the background technology, the purpose of the embodiments of the present invention is to provide a method, device and storage medium for generating technical interview questions based on AIAgent to solve the following technical problems:

[0022] (1) Automated and efficient processing: The recruitment process involving manual participation has been greatly reduced, thus improving recruitment efficiency.

[0023] (2) Accurate decomposition of technical labels: AI Agent accurately decomposes technical labels to ensure comprehensive coverage of job requirements.

[0024] (3) Highly targeted: The generated interview questions can be personalized according to the candidate’s specific background, which improves the effectiveness of the interview.

[0025] (4) Dynamic adaptability: The system can automatically adjust technical tags and interview questions based on new job requirements and market changes to keep the recruitment process up to date.

[0026] To achieve the above objectives, in a first aspect, an embodiment of the present invention provides a method for generating technical interview questions based on an AI Agent, comprising:

[0027] Enter the job description information and candidate's resume information;

[0028] Use natural language processing technology and AI Agent to process the job description information and resume information to generate original technical labels;

[0029] Based on the project information in the job description information, use AI Agent to infer the project technical architecture that the job may involve;

[0030] Generate a full set of technical tags based on the original technical tags and the project technical architecture;

[0031] Automatically generate technical interview questions based on all technical tags, resume information and interview question bank.

[0032] As a specific implementation method, the original technical label is generated as follows:

[0033] Using natural language processing technology to extract technical keywords from the job description information and resume information;

[0034] AI Agent is used to perform in-depth analysis on the technical keywords to generate a set of original technical tags related to the original job requirements.

[0035] As a specific implementation method, the technical architecture inferred by AI Agent is as follows:

[0036] The project information in the job description information inputted in combination with domain knowledge and historical recruitment data is analyzed by AIAgent to obtain the project technical architecture that may be involved in the job; the project technical architecture includes the programming language, framework and database used.

[0037] As a specific implementation method, generating a full technical label according to the original technical label and technical architecture is specifically as follows:

[0038] Based on the original technical labels, project technical architecture and key technical details in the project, AIAgent is used to perform reasoning to obtain the project's technical stack and corresponding knowledge points;

[0039] Use AI Agent to evaluate the difficulty of the knowledge points and divide them into three levels: "easy", "medium" and "difficult";

[0040] Using AI Agent to infer the knowledge points after difficulty evaluation to obtain derived labels;

[0041] The derived label and the original technical label are combined to obtain a full technical label.

[0042] As a specific implementation method, the technical interview questions are automatically generated based on the full technical tags, resume information and interview question bank as follows:

[0043] Taking the plurality of said full technical tags as an input technical tag list, and using the AI ​​Agent to screen each full technical tag to filter out full technical tags with low weights;

[0044] According to the difficulty level of the screened technical tags and knowledge points, technical interview questions and reference answers are selected from the interview question bank;

[0045] Create an audit AI Agent, and use the audit AI Agent to audit the technical interview questions and reference answers;

[0046] Output the technical interview questions and reference answers that have passed the review.

[0047] Further, as a preferred implementation of the present application, after generating the technical interview questions, the method further includes interview question feedback, specifically:

[0048] Collect feedback data from interviewers and candidates on interview questions;

[0049] Based on the feedback data, perform question effect analysis, difficulty analysis and adjustment on the interview questions, and correlation analysis between the questions and job requirements to obtain analysis results;

[0050] Adjust the interview question bank and technical label generation strategy based on the analysis results.

[0051] In a second aspect, an embodiment of the present application provides a device for generating technical interview questions based on an AI Agent, comprising:

[0052] Input unit, used to input job description information and candidate resume information;

[0053] An AI Agent unit is used to process the job description information and resume information using natural language processing technology and AI Agent to generate original technical labels;

[0054] The AI ​​Agent unit is further used to use the AI ​​Agent to infer the project technology architecture that may be involved in the position based on the project information in the position description information;

[0055] A technical label expansion unit, used to generate a full technical label according to the original technical label and the project technical architecture;

[0056] Interview question generation unit, used to automatically generate technical interview questions based on all technical tags, resume information and interview question bank;

[0057] The output unit is used to push the technical interview questions to the interviewer or the candidate.

[0058] In the third aspect, an embodiment of the present invention further provides an AI Agent-based technical interview question generation device, comprising a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, and the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method of the first aspect above.

[0059] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium storing a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the method of the first aspect.

[0060] The implementation of the embodiment of the present invention has the following advantages:

[0061] (1) AI Agent is used to intelligently decompose and expand the technical labels of job descriptions (JDs) and infer complex technical architectures. Through natural language processing (NLP) and machine learning algorithms, the system can automatically generate technical labels that are highly matched with job requirements and dynamically expand these labels to cover all core skill requirements of the job, which can significantly improve the accuracy and efficiency of the recruitment process.

[0062] (2) Based on the job description, project background, company background, and applicant resume information, the system automatically generates technical interview questions that match the technical tags. These interview questions not only reflect the skills required for the position, but can also be personalized based on the applicant's background to ensure the relevance and pertinence of the interview questions.

[0063] (3) Dynamic feedback and optimization mechanism: The system has the ability to dynamically provide feedback and optimize the recruitment process based on the interview results. By analyzing the feedback from interviewers and candidates, the system can automatically adjust the interview question bank and label generation strategy to continuously improve the effectiveness and accuracy of the interview.

[0064] (4) Efficient recruitment process automation: The system greatly reduces the number of manual processes involved in the recruitment process through intelligent technical label decomposition, technical architecture inference, label generation, and interview question generation, significantly improving recruitment efficiency and accuracy. The automation of this process significantly reduces human resource management costs while improving recruitment quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the drawings required for use in the specific implementation or the description of the prior art are briefly introduced below.

[0066] Figure 1 is a flow chart of a method for generating technical interview questions based on an AI Agent provided in an embodiment of the present invention;

[0067] Figure 2 It is a flowchart of data preprocessing for interview question generation;

[0068] Figure 3 It is a flowchart for generating technical interview questions;

[0069] Figure 4 It is a technical interview feedback flow chart;

[0070] Figure 5 It is a technical interview question optimization flow chart;

[0071] Figure 6 It is a structural diagram of a device for generating technical interview questions based on AI Agent provided in an embodiment of the present invention;

[0072] Figure 7 yes Figure 6 Another structural diagram of the device shown. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0075] JD: JD is the abbreviation of Job Description, which refers to a job description or position specification. It is usually used in the recruitment process and is a document that lists the job requirements, responsibilities, necessary skills, experience and qualifications in detail.

[0076] Agent: Agent in the field of artificial intelligence usually refers to an "intelligent body", that is, an autonomous system or program that can perceive the environment and take actions to achieve specific goals. In smart recruitment, Agent can refer to an automated AI system that can autonomously make decisions or generate outputs based on input data.

[0077] Large model: Large model usually refers to Large Language Model, especially in the field of Natural Language Processing (NLP). Large model is a deep learning model trained based on a large amount of text data, usually with billions or even hundreds of billions of parameters.

[0078] Intelligent Recruiting: Intelligent Recruiting is a method that uses artificial intelligence (AI) technology to optimize and automate the recruitment process. It helps companies screen, evaluate and select candidates more efficiently through AI technologies such as natural language processing, machine learning, and data analysis.

[0079] The inventive concept of the present invention is to provide a method for generating technical interview questions for an intelligent recruitment system based on AI Agent, decompose the job description into technical tags through AI Agent, and infer the project technical architecture required for the job based on these technical tags. Finally, through further expansion of tags and association with the interview question bank, customized interview questions for candidates are generated, thereby improving recruitment efficiency and accuracy.

[0080] Please refer to Figure 1 The present invention provides a method for generating technical interview questions based on AI Agent, comprising:

[0081] S1, input stage: input job description information and candidate resume information.

[0082] S2, technical label decomposition: use natural language processing technology and AI Agent to process the job description information and resume information to generate original technical labels.

[0083] Specifically, the system first receives the input job description (JD) and the candidate's resume information, and uses natural language processing technology to extract core technical keywords. Then, the AI ​​Agent performs in-depth analysis of these keywords to generate a set of original technical tags related to the job requirements.

[0084] S3, project technical architecture inference: Based on the project information in the job description information, use AI Agent to infer the project technical architecture that the job may involve.

[0085] Specifically, the project information in the job description information is combined with domain knowledge and historical recruitment data input, and the project technical architecture that may be involved in the job is inferred through AI Agent and large models; the project technical architecture includes the programming language, framework and database used.

[0086] S4, technical label generation and expansion: Generate a full set of technical labels based on the original technical labels and project technical architecture.

[0087] After inferring the project's technical architecture, the system will further generate related technical tags and expand these tags to cover all key technical areas required for the position. This step ensures that the generated technical tags can fully and accurately describe the job requirements.

[0088] The specific process of this step is as follows:

[0089] (1) Based on the original technical labels, project technical architecture, and key technical details of the project, the big model is used to infer the project's technology stack and corresponding knowledge points.

[0090] (2) The AI ​​Agent evaluates the difficulty of knowledge points and divides them into three levels: "easy", "medium", and "difficult".

[0091] (3) Use AI Agent to further reason about all knowledge points to obtain more technical labels, and merge these derived labels with the original technical labels to obtain the full set of technical labels.

[0092] S5, interview question generation: automatically generate technical interview questions based on all technical tags, resume information and interview question bank.

[0093] The interview question bank is generated based on the original question bank. The detailed process is as follows: Figure 2 As shown, it includes: interview question crawling → interview question decomposition → job label generation → difficulty label generation → technical label generation → storing the corresponding relationship in the database.

[0094] The system combines the generated technical tags, the applicant's resume information and the interview question bank to automatically generate a set of targeted technical interview questions. These questions cover the core skills required for the position and can be personalized based on the candidate's background. Figure 3 As shown in the figure, the process of generating technical interview questions includes technical tag list → weight analysis → finding corresponding interview questions → large model imitates and generates similar interview questions → verifies the interview questions generated by the large model → outputs complete interview questions, specifically:

[0095] (1) All technical tags are used as the input technical tag list, and an AI agent analyzes the weight of each tag and filters out technical tags with low weight.

[0096] (2) Filter out appropriate interview questions from the question bank based on the screened technical tags and difficulty levels.

[0097] (3) AI Agent generates a similar interview question and reference answer based on the technical tags and interview questions.

[0098] Among them, interview questions can be divided into: single-choice questions, multiple-choice questions, fill-in-the-blank questions, judgment questions, and question-and-answer questions.

[0099] (4) Create another AI Agent as a reviewer of the interview questions to determine whether the question and answer are correct.

[0100] (5) Output the approved interview questions and reference answers to the next module.

[0101] S6, Interview question feedback: Collect feedback data from interviewers and candidates on interview questions, and analyze them to adjust the interview question bank and technical label generation strategy.

[0102] like Figure 4 As shown in the figure, the interview question feedback process mainly includes obtaining the interview position → obtaining the position level → labeling the interview questions → generating a reinforcement learning dataset.

[0103] Please refer to Figure 5 ,The technical interview question optimization process includes collecting feedback data → data analysis and processing → optimizing the interview question bank → feedback-driven algorithm optimization → real-time improvement and monitoring → loop iteration.

[0104] 1. Collect feedback data

[0105] Feedback data mainly comes from interviewers and candidates.

[0106] Among them, interviewer feedback: collect interviewers' evaluations of interview questions, including the difficulty and relevance of the questions, the candidate's performance, and whether the questions effectively assess the candidate's skills.

[0107] Candidate feedback: After the interview, you can distribute questionnaires to candidates or conduct a brief interview experience survey to collect their opinions on the questions, such as whether the questions are clear, whether the difficulty is appropriate, and whether the questions reflect the job requirements.

[0108] 2. Data Analysis and Processing

[0109] Data analysis and processing mainly include question effect analysis, difficulty adjustment and correlation analysis.

[0110] (1) Question effectiveness analysis: Through data analysis, identify which interview questions show a high degree of discrimination when evaluating candidates, that is, these questions can effectively distinguish candidates with different skill levels. These questions should be retained and strengthened.

[0111] Among them, the algorithm for evaluating interview questions is as follows:

[0112] ①Data Collection

[0113] Input: Collect scores of multiple candidates on a set of interview questions. Suppose we have a list of candidates, and the scores of each candidate on each question form a matrix, scores, where scores[i][j] represents the score of candidate i on interview question j.

[0114] Label: The final interview result of each candidate (pass / fail, or other qualitative evaluation) is also used as a label vector labels.

[0115] ②Calculate the discrimination

[0116] Score difference: For each interview question j, calculate the score difference between the high-scoring candidate group and the low-scoring candidate group on this question. This score difference reflects the degree of discrimination between candidates of different levels.

[0117] Variance analysis: Calculate the variance of the scores of candidates on each question. The larger the variance, the greater the difference in performance of different candidates on the question, and the higher the discrimination may be.

[0118] Here are the specific steps:

[0119] Candidate ranking: Candidates are ranked according to their total scores.

[0120] Grouping: Candidates are divided into a high group (top 25%) and a low group (bottom 25%).

[0121] Difference calculation: For each interview question j, calculate the difference in the average scores between the high-scoring group and the low-scoring group, denoted as D j .

[0122] Variance calculation: Calculate the score variance Var for each question j .

[0123] ③ Comprehensive discrimination calculation

[0124] The score difference D j and score variance Var j Combined, we can define the comprehensive discrimination S of each question j .

[0125] S j =α·D j +β·sqrt(Var j )

[0126] Among them, α and β are adjustment parameters used to balance the impact of score difference and variance.

[0127] ④Evaluation results

[0128] Discrimination sorting: according to S j Sort the interview questions, S j The higher the discrimination of the questions, the better.

[0129] Screening: Select the questions with the highest discrimination as the final screening questions.

[0130] (2) Difficulty Adjustment: Analyze the difficulty of each question based on candidate feedback and performance data. If the pass rate of a question is too low or too high, it may be necessary to adjust its difficulty or redesign the question to ensure it is appropriately challenging.

[0131] (3) Relevance analysis: Evaluate the relevance of each question to the job requirements. If some questions are weakly related to the actual job requirements, you can consider replacing or redesigning them. The detailed steps are as follows:

[0132] ①Prepare data

[0133] a) Objective: To evaluate the relevance of each interview question to the job requirements.

[0134] b) Input:

[0135] ●Job description: including job responsibilities, skill requirements, etc.

[0136] ●Project architecture information: describes the project’s technical architecture, modules, etc.

[0137] ●Technology stack: the specific technologies required for the position, such as programming languages, frameworks, tools, etc.

[0138] ●Interview question bank: contains multiple interview questions related to the position.

[0139] c) Output: The relevance score of each interview question to the job requirements, indicating the extent to which it covers the job requirements.

[0140] ②Analyze input data

[0141] Job description: usually contains some keywords, such as "proficient in Python programming", "proficient in database design", etc. These keywords can serve as an important basis for relevance assessment.

[0142] Project architecture information: describes the composition and interrelationships of the various modules of the system, which can help understand the role of this position in the project.

[0143] Technology stack: Specific list of technologies required for the position, which will be directly used to evaluate the relevance of interview questions.

[0144] Interview question bank: Each question can contain information such as the question stem, difficulty, and knowledge points involved.

[0145] ③Build key features

[0146] a) Job feature extraction:

[0147] Keyword extraction: Extract keywords from job descriptions, project architecture information, and technology stacks. These keywords can include programming languages, tools, frameworks, technology fields, etc.

[0148] Weight assignment: Different keywords may have different importance in the job description. You can assign weights to keywords manually or through historical data.

[0149] b) Interview question feature extraction:

[0150] Keyword matching: Analyze each interview question and extract the technical keywords involved in the question stem.

[0151] Content analysis: Analyze the content of the interview questions to assess the technical scope and depth of the questions.

[0152] ④Calculate correlation

[0153] a) Keyword matching:

[0154] Direct Match: Calculates the direct match between the job description and the interview question keywords. The more keywords matched, the stronger the correlation.

[0155] Synonym Expansion: Consider synonyms in the technology field and improve accuracy by expanding matching.

[0156] b) Technology stack coverage:

[0157] Technology coverage: Calculate the overlap between the technology stack involved in the interview questions and the job requirements. The higher the overlap, the stronger the correlation.

[0158] c) Project architecture relevance:

[0159] Module Match: If the interview question involves a specific project module, consider the relevance of that module to the position.

[0160] ⑤Design algorithm

[0161] Relevance scoring formula:

[0162] Relevance score = α × keyword matching score + β × technology stack coverage score + γ × project architecture relevance score

[0163] Among them, α, β and γ are weight parameters for adjusting the contribution of each part, which can be adjusted according to the actual application.

[0164] Keyword matching score: Keyword matching score = number of matched keywords / total number of job keywords

[0165] Technology stack coverage score: Technology stack coverage score = number of matched technology stack items / total number of technology stack items

[0166] Project architecture relevance score: Project architecture relevance score = number of matched project modules / total number of project modules

[0167] ⑥Optimization and verification

[0168] Dataset construction: Use historical data (such as the correlation between previous interview questions and candidate performance) to train and verify the algorithm, and optimize the algorithm's effect by adjusting the weight parameters α, β, and γ.

[0169] Model validation: Evaluate the performance of the algorithm on different data sets through methods such as cross-validation and holdout validation, and continuously iterate and optimize.

[0170] (III) Optimize the interview question bank

[0171] Question bank update: Based on the analysis results, the existing interview question bank is updated. Poorly performing questions can be deleted and new, more relevant and effective questions can be added.

[0172] Personalized Adjustment: Introducing dynamically adjustable questions in the question bank, such as adjusting the difficulty or direction of subsequent questions based on the candidate’s answer performance. In this way, the interview process can be more personalized and the candidate’s skills can be assessed more accurately.

[0173] 4. Feedback-driven algorithm improvement

[0174] Machine learning model optimization: Use the collected feedback data to further train and optimize the AI ​​model that generates interview questions. By incorporating feedback data into the model training process, AI can learn which questions work best and generate better interview questions in the future.

[0175] Reinforcement learning application: Introduce reinforcement learning algorithms to adjust the strategy for generating interview questions in real time. For example, the system can dynamically update the strategy for generating interview questions based on interview feedback, making the questions more suitable for job requirements.

[0176] (V) Implementation, improvement and monitoring

[0177] Experimentation and testing: After updating and optimizing the interview question bank, you can conduct a small-scale test to ensure that the optimized question bank performs well in actual applications. If necessary, further adjustments and optimizations can be made.

[0178] Continuous monitoring: While improving the interview question bank, continuously monitor the feedback data during the interview process to ensure that the optimization measures are effective and the interview question bank continues to meet job requirements and candidate performance.

[0179] (VI) Loop Iteration

[0180] Circular optimization: The optimization of interview questions is a continuous cycle. The feedback data collected after each interview should be included in the analysis to provide a reference for the generation of the next interview questions. Through cyclic iteration, the interview question bank will gradually mature and optimize.

[0181] It can be seen from the above description that the method provided by the embodiment of the present invention has the following advantages:

[0182] (1) AI Agent is used to intelligently decompose and expand the technical labels of job descriptions (JDs) and infer complex technical architectures. Through natural language processing (NLP) and machine learning algorithms, the system can automatically generate technical labels that are highly matched with job requirements and dynamically expand these labels to cover all core skill requirements of the job, which can significantly improve the accuracy and efficiency of the recruitment process.

[0183] (2) Based on the job description, project background, company background, and applicant resume information, the system automatically generates technical interview questions that match the technical tags. These interview questions not only reflect the skills required for the position, but can also be personalized based on the applicant's background to ensure the relevance and pertinence of the interview questions.

[0184] (3) Dynamic feedback and optimization mechanism: The system has the ability to dynamically provide feedback and optimize the recruitment process based on the interview results. By analyzing the feedback from interviewers and candidates, the system can automatically adjust the interview question bank and label generation strategy to continuously improve the effectiveness and accuracy of the interview.

[0185] (4) Efficient recruitment process automation: The system greatly reduces the number of manual processes involved in the recruitment process through intelligent technical label decomposition, technical architecture inference, label generation, and interview question generation, significantly improving recruitment efficiency and accuracy. The automation of this process significantly reduces human resource management costs while improving recruitment quality.

[0186] Based on the same inventive concept, Figure 6 As shown, the embodiment of the present invention also provides a device for generating technical interview questions based on AI Agent, including:

[0187] Input unit, used to input job description information and candidate resume information;

[0188] An AI Agent unit is used to process the job description information and resume information using natural language processing technology and AI Agent to generate original technical labels;

[0189] The AI ​​Agent unit is further used to use the AI ​​Agent to infer the project technology architecture that may be involved in the position based on the project information in the position description information;

[0190] A technical label expansion unit, used to generate a full technical label according to the original technical label and the project technical architecture;

[0191] Interview question generation unit, used to automatically generate technical interview questions based on all technical tags, resume information and interview question bank;

[0192] An output unit, used to push the technical interview questions to the interviewer or the candidate;

[0193] The interview feedback unit is used to collect feedback data and analyze it to adjust the interview question bank and technical label generation strategy.

[0194] Among them, the original technical labels generated by the AI ​​Agent unit are as follows:

[0195] Using natural language processing technology to extract technical keywords from the job description information and resume information;

[0196] AI Agent is used to perform in-depth analysis on the technical keywords to generate a set of original technical tags related to the original job requirements.

[0197] Furthermore, the AI ​​Agent unit deduces that the technical architecture is as follows:

[0198] The project information in the job description information inputted in combination with domain knowledge and historical recruitment data is analyzed by AIAgent to obtain the project technical architecture that may be involved in the job; the project technical architecture includes the programming language, framework and database used.

[0199] Furthermore, the technical label extension unit generates the full technical label specifically as follows:

[0200] Based on the original technical labels, project technical architecture and key technical details in the project, AIAgent is used to perform reasoning to obtain the project's technical stack and corresponding knowledge points;

[0201] Use AI Agent to evaluate the difficulty of the knowledge points and divide them into three levels: "easy", "medium" and "difficult";

[0202] Using AI Agent to infer the knowledge points after difficulty evaluation to obtain derived labels;

[0203] The derived label and the original technical label are combined to obtain a full technical label.

[0204] Furthermore, the interview question generation unit automatically generates technical interview questions specifically as follows:

[0205] Taking the plurality of said full technical tags as an input technical tag list, and using the AI ​​Agent to screen each full technical tag to filter out full technical tags with low weights;

[0206] According to the difficulty level of the screened technical tags and knowledge points, technical interview questions and reference answers are selected from the interview question bank;

[0207] Create an audit AI Agent, and use the audit AI Agent to audit the technical interview questions and reference answers;

[0208] Output the technical interview questions and reference answers that have passed the review.

[0209] Furthermore, the interview feedback unit is specifically used to:

[0210] Collect feedback data from interviewers and candidates on interview questions;

[0211] Based on the feedback data, perform question effect analysis, difficulty analysis and adjustment on the interview questions, and correlation analysis between the questions and job requirements to obtain analysis results;

[0212] Adjust the interview question bank and technical label generation strategy based on the analysis results.

[0213] It should be noted that the specific working process of this embodiment can be found in the aforementioned method embodiment part, which will not be described in detail here.

[0214] Furthermore, another embodiment of the present invention provides a device for generating technical interview questions based on AI Agent. Figure 7 As shown, the apparatus may include: one or more processors 101, one or more input devices 102, one or more output devices 103 and a memory 104, wherein the processors 101, the input devices 102, the output devices 103 and the memory 104 are interconnected via a bus 105. The memory 104 is used to store a computer program, wherein the computer program includes program instructions, and the processor 101 is configured to call the program instructions to execute the method of the above method embodiment.

[0215] It should be understood that in the embodiment of the present invention, the processor 101 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0216] The input device 102 may include a keyboard, etc., and the output device 103 may include a display (LCD, etc.), a speaker, etc.

[0217] The memory 104 may include a read-only memory and a random access memory, and provides instructions and data to the processor 101. A portion of the memory 104 may also include a non-volatile random access memory. For example, the memory 104 may also store information on the device type.

[0218] In a specific implementation, the processor 101, input device 102, and output device 103 described in the embodiment of the present invention can execute the implementation method described in the embodiment of the method for generating technical interview questions based on AI Agent provided in the embodiment of the present invention, which will not be repeated here.

[0219] Accordingly, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the above-mentioned method for generating technical interview questions based on AI Agent is implemented.

[0220] The computer-readable storage medium may be an internal storage unit of the system described in any of the foregoing embodiments, such as a hard disk or memory of the system. The computer-readable storage medium may also be an external storage device of the system, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the system. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the system and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the system. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0221] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0222] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or it can be an electrical, mechanical or other form of connection.

[0223] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0224] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0225] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0226] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for generating technical interview questions based on AIAgent, characterized in that: include: Enter the job description information and candidate's resume information; Use natural language processing technology and AI Agent to process the job description information and resume information to generate original technical labels; Based on the project information in the job description information, use AI Agent to infer the project technical architecture that the job may involve; Generate a full set of technical tags based on the original technical tags and the project technical architecture; Automatically generate technical interview questions based on all technical tags, resume information and interview question bank.

2. The method according to claim 1, characterized in that Generate the original technical label as follows: Using natural language processing technology to extract technical keywords from the job description information and resume information; AI Agent is used to perform in-depth analysis on the technical keywords to generate a set of original technical tags related to the original job requirements.

3. The method according to claim 1, characterized in that The specific technical architecture inferred by AI Agent is as follows: The project information in the job description information inputted in combination with domain knowledge and historical recruitment data is analyzed through AI Agent to obtain the project technical architecture that may be involved in the job; the project technical architecture includes the programming language, framework and database used.

4. The method according to claim 1, characterized in that The full technical labels generated according to the original technical labels and technical architecture are as follows: Based on the original technical labels, project technical architecture, and key technical details in the project, use AI Agent to perform reasoning to obtain the project's technical stack and corresponding knowledge points; Use AI Agent to evaluate the difficulty of the knowledge points and divide them into three levels: "easy", "medium" and "difficult"; Using AIAgent to infer the knowledge points after difficulty evaluation to obtain derived labels; The derived label and the original technical label are combined to obtain a full technical label.

5. The method according to claim 2, characterized in that The technical interview questions are automatically generated based on the full technical tags, resume information and interview question bank: Taking the plurality of said full technical tags as an input technical tag list, and using the AI ​​Agent to screen each full technical tag to filter out full technical tags with low weights; According to the difficulty level of the screened technical tags and knowledge points, technical interview questions and reference answers are selected from the interview question bank; Create an audit AI Agent, and use the audit AI Agent to audit the technical interview questions and reference answers; Output the technical interview questions and reference answers that have passed the review.

6. The method according to any one of claims 1 to 5, characterized in that: After generating the technical interview questions, the method further includes providing interview question feedback, specifically: Collect feedback data from interviewers and candidates on interview questions; Based on the feedback data, perform question effect analysis, difficulty analysis and adjustment on the interview questions, and correlation analysis between the questions and job requirements to obtain analysis results; Adjust the interview question bank and technical label generation strategy based on the analysis results.

7. A device for generating technical interview questions based on AIAgent, characterized in that: include: Input unit, used to input job description information and candidate resume information; An AI Agent unit is used to process the job description information and resume information using natural language processing technology and AI Agent to generate original technical labels; The AI ​​Agent unit is further used to use the AI ​​Agent to infer the project technology architecture that may be involved in the position based on the project information in the position description information; A technical label expansion unit, used to generate a full technical label according to the original technical label and the project technical architecture; Interview question generation unit, used to automatically generate technical interview questions based on all technical tags, resume information and interview question bank; The output unit is used to push the technical interview questions to the interviewer or the candidate.

8. The device according to claim 7, characterized in that The AI ​​Agent unit is specifically used for: Using natural language processing technology to extract technical keywords from the job description information and resume information; Use AI Agent to deeply analyze the technical keywords and generate a set of original technical tags related to the original job requirements; The project information in the job description information inputted in combination with domain knowledge and historical recruitment data is analyzed through AI Agent to obtain the project technical architecture that may be involved in the job; the project technical architecture includes the programming language, framework and database used.

9. A device for generating technical interview questions based on AI Agent, characterized in that: The method comprises a processor, an input device, an output device and a memory, wherein the processor, the input device, the output device and the memory are interconnected, wherein the memory is used to store a computer program, the computer program comprises program instructions, and the processor is configured to call the program instructions to execute the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor is caused to perform the method according to any one of claims 1 to 6.

Images