Artificial intelligence-based intelligent matching and dynamic scheduling system and method for IT personnel outsourcing

Abstract

This invention relates to the field of computer software technology and discloses an AI-based intelligent matching and dynamic scheduling system and method for IT personnel outsourcing. The system includes a data acquisition module, a feature engineering module, a knowledge graph module, an intelligent matching engine module, a dynamic scheduling module, an effect prediction module, and a performance evaluation module. By collecting and processing personnel information, project requirement data, and business data, it constructs personnel capability feature vectors and project requirement feature vectors, and combines these with personnel skill graphs, project experience graphs, technical difficulty graphs, and solution graphs to achieve knowledge association analysis. The system optimizes the allocation of IT personnel to project requirements through intelligent matching and dynamic scheduling, and continuously optimizes the matching results through effect prediction and performance evaluation. This invention can improve the efficiency of IT outsourcing personnel allocation and the accuracy of project matching, thereby increasing project execution success rate and customer satisfaction.

Description

Technical Field

[0001] This invention relates to the field of computer software technology, and specifically to an intelligent matching and dynamic scheduling system and method for IT personnel outsourcing based on artificial intelligence. Background Technology

[0002] With the accelerating pace of global digital transformation, IT talent outsourcing has become a crucial channel for enterprises to acquire key technological capabilities. IT personnel outsourcing refers to companies contracting out certain IT-related tasks or positions to professional human resource service agencies, which are then responsible for recruitment, training, management, and deployment. However, traditional IT personnel outsourcing models suffer from the following significant problems: I. Inefficient Person-Job Matching. Traditional outsourced staffing relies primarily on manual resume screening and interview assessments, a time-consuming and subjective process. HR personnel must sift through a large number of resumes one by one, which is inefficient and prone to overlooking excellent candidates. Furthermore, manual assessments struggle to comprehensively quantify candidates' technical skills, project experience, teamwork abilities, and other key factors, resulting in inconsistent matching quality.

[0003] Second, there is a lack of flexibility in personnel scheduling. During project execution, requirements change frequently, and static personnel allocation plans are difficult to adapt to dynamic project needs. When project delays, staff leave, or technical difficulties arise, it is impossible to quickly adjust personnel and reallocate resources, affecting project progress and quality.

[0004] Third, there is a lack of effective means of predicting outcomes. Traditional management models can only respond passively after problems occur, and cannot predict project risks and the effectiveness of staffing in advance. Companies lack data support when selecting outsourced personnel, making it difficult to make optimal decisions.

[0005] Fourth, personnel performance evaluation is not comprehensive. Existing performance evaluations often only focus on project delivery results, neglecting a comprehensive evaluation of multiple dimensions such as personnel technical growth, team cohesion, and customer satisfaction, thus failing to provide valuable reference for personnel allocation in subsequent projects.

[0006] Fifth, difficulties in knowledge transfer. Frequent turnover of outsourced personnel makes it difficult to effectively accumulate and transfer project experience and knowledge. Each time a new person enters a project, they need to rebuild their understanding, which increases management costs and project risks.

[0007] Given the aforementioned shortcomings, there is an urgent need for a new technical solution to address issues such as job matching, personnel scheduling, performance prediction, and performance evaluation in the IT personnel outsourcing field. Summary of the Invention

[0008] To achieve the above-mentioned objectives, the present invention provides the following technical solution: an intelligent matching and dynamic scheduling system for IT personnel outsourcing based on artificial intelligence, comprising the following modules: The data acquisition module is used to obtain IT personnel data, project requirement data, and business process data from multiple business systems, and to perform unified aggregation and standardized processing on the collected data. The feature engineering module is used to perform structured processing and feature vector construction on the collected data. It generates unified personnel capability feature vectors and project requirement feature vectors through text feature extraction, numerical feature processing, and category feature encoding. The intelligent matching engine module is used to screen, match, sort, and generate a recommendation list of candidates based on project requirements and personnel capabilities. The dynamic scheduling module is used to monitor project progress and personnel status in real time during project execution, and generate dynamic adjustment plans and execute personnel scheduling when changes are detected. The effect prediction module is used to predict and analyze the project success rate, staff retention rate and customer satisfaction after the matching plan is implemented, in order to evaluate the feasibility of the current matching plan. The performance evaluation module is used to evaluate the comprehensive performance of IT personnel based on project execution results and multi-dimensional evaluation data, and to generate performance evaluation results for updating personnel capability profiles. The knowledge graph module is used to build a knowledge graph in the field of IT personnel outsourcing. By associating and modeling personnel skill information, project experience information, technical difficulty information, and solution information, it forms personnel skill graphs, project experience graphs, technical difficulty graphs, and solution graphs. These graphs are then stored and analyzed through a graph database to provide knowledge support for personnel matching, scheduling decisions, and effect prediction.

[0009] The data acquisition module includes: The data source access submodule is used to establish data interface connections with the internal human resources system, project management system, customer feedback system, attendance system, and training system to obtain basic personnel information, project contract information, project execution data, customer evaluation information, and training record information. The data synchronization submodule is used to periodically synchronize and update data in various business systems using a data synchronization strategy that combines incremental synchronization and full backup, and to ensure the timeliness and consistency of data updates through timestamp and version identification mechanisms. The data preprocessing submodule is used to clean, deduplicate, convert formats, and impute missing values ​​in the collected data, thereby forming structured and unstructured datasets. The processed data is then stored in a unified data warehouse for subsequent modules to use.

[0010] The feature engineering module includes: The text feature extraction submodule is used to perform semantic analysis on unstructured text data, including resume text, project report and technical document, and extracts semantic vectors based on BERT pre-trained language model to generate text semantic feature vectors. The numerical feature processing submodule is used to perform normalization and binning processing on numerical data, including years of work experience, number of projects completed, and customer ratings, in order to eliminate differences between data of different dimensions. The category feature encoding submodule is used to perform one-hot encoding on categorical data, which includes technology stack categories, job categories and project types, thereby generating discrete feature representations that can be used for model training. The temporal feature extraction submodule is used to extract features from personnel's historical project participation records and skill growth trajectories using a sliding window method to construct personnel capability change trend features and dynamically update personnel feature vectors.

[0011] The intelligent matching engine module includes: The candidate screening submodule is used to initially screen candidates based on the technical stack requirements, work experience requirements, education requirements, and certification requirements in the project requirements, and to quickly locate the set of candidates that meet the conditions in the candidate database through an inverted index structure. The ranking optimization submodule is used to comprehensively rank the candidates who have passed the initial screening. It obtains the matching score by calculating the cosine similarity between the personnel's ability feature vector and the project's requirement feature vector. At the same time, it constructs a multi-objective optimization model by combining the personnel's current workload index, historical cooperation effect index, and customer historical preference index, and calculates the candidate's comprehensive score through a weighted summation model. The diversity preservation submodule is used to evaluate the diversity of the ranking results. It judges the concentration of the recommendation results by calculating the feature similarity between recommenders. When the inter-group difference of the recommendation results is lower than a set threshold, the recommendation list is readjusted by introducing differentiated candidates to improve the diversity of the recommendation results.

[0012] The dynamic scheduling module includes: The anomaly detection submodule is used to monitor key indicators in real time during project execution, including project progress indicators, personnel attendance indicators, and work output indicators, and to identify abnormal events and trigger early warning mechanisms based on time-series data anomaly detection algorithms. The impact assessment submodule is used to assess the scope of impact of change factors when a project change request is received, and to comprehensively analyze the technical complexity of the change, the time urgency, and the intensity of resource requirements, thereby generating the corresponding impact score and risk level. The scheme generation submodule is used to optimize and search for personnel allocation schemes based on reinforcement learning strategies. It constructs a reward function to comprehensively evaluate project completion rate, personnel utilization rate and customer satisfaction, and selects the optimal scheduling scheme from multiple candidate schemes. The automatic execution submodule is used to execute the scheduling plan after obtaining management authorization, including personnel allocation notification, project plan update and data synchronization update of related business systems, and to monitor the execution status in real time to ensure the stability of the scheduling process.

[0013] The effect prediction module includes: The project success rate prediction submodule is used to build a multilayer perceptron prediction model. It takes the matching scheme features as the model input, trains it with historical project data, and outputs the project success probability. The training data includes project features, personnel configuration features, and project result labels. The model is optimized using the cross-entropy loss function and uses dropout technology to prevent overfitting. The employee retention rate prediction submodule is used to predict the probability of employees continuing to work in a project based on a survival analysis model, and to generate short-term retention probability and long-term retention probability by comprehensively analyzing factors such as employee job satisfaction, team cohesion, technical challenges and salary competitiveness. The customer satisfaction prediction submodule is used to build a regression prediction model to predict customer satisfaction ratings for project outcomes. Its inputs include project characteristics, staffing characteristics, and project execution process characteristics, and its output is the predicted satisfaction rating. The three prediction results together form a project performance prediction report. When the project success probability is lower than a preset threshold, the system triggers a re-matching mechanism.

[0014] The performance evaluation module includes: The performance indicator construction submodule is used to establish a multi-dimensional performance indicator system, including project delivery quality indicators, work attitude indicators, team collaboration indicators, technical growth indicators, and customer feedback indicators. The evaluation data collection submodule is used to collect self-evaluation data, supervisor evaluation data, peer evaluation data, and customer evaluation data, and to normalize them using a unified scoring standard. The performance analysis submodule is used to comprehensively calculate the scores of each dimension using a 360-degree evaluation method, and generate a performance radar chart and performance details table in a visual way to form a comprehensive performance evaluation result for IT personnel. The evaluation result is then fed back to the personnel capability profile to support subsequent matching and optimization.

[0015] The knowledge graph module includes: The Personnel Skills Mapping submodule is used to build a personnel skills knowledge network with IT personnel as the core node, record the technical skills, skill levels and skill development trajectories of personnel, and establish the relationship between personnel and technical skills; The Project Experience Graph submodule is used to build an experience knowledge network of personnel participating in projects, record project types, technology stack structure, personnel responsibilities and roles, and project deliverables, and establish a mapping between personnel and projects; The Technical Challenge Map submodule is used to record the technical challenges encountered during project implementation and to establish the relationship between technical challenges and technical fields, project types, and related technology stacks. The Solution Mapping submodule is used to record solutions, handling strategies, and best practice experiences for technical challenges, and to establish the relationship between solutions and technical challenges. The knowledge graph is stored in the Neo4j graph database and knowledge nodes and relationship edges are constructed through a combination of automatic information extraction and manual review to support personnel matching, technical reasoning and knowledge retrieval.

[0016] This invention also provides an AI-based intelligent matching and dynamic scheduling method for IT personnel outsourcing, comprising the following steps: Step S1: Collect basic information of IT personnel, project requirements, project execution data and customer evaluation data from the human resources system, project management system, customer feedback system, attendance system and training system through the data acquisition module, and store the data in a unified data warehouse after data cleaning, format unification, missing value filling and standardization. Step S2: The feature engineering module performs feature extraction and feature encoding on the collected data. Semantic vectors are extracted from resume text, project reports and technical documents. Work experience years, number of completed projects and customer rating data are normalized. Technology stack category, job category and project type are encoded and corresponding personnel ability feature vectors and project requirement feature vectors are generated. Step S3: Extract knowledge and construct relationships from IT personnel skill information, project experience information, technical difficulty information, and solution information through the knowledge graph module to form a knowledge network that includes personnel skill graph, project experience graph, technical difficulty graph, and solution graph. Obtain personnel skill association information and project experience association information based on graph query and relational reasoning. Step S4: The intelligent matching engine module filters candidates based on the project requirement feature vector and the personnel skill association features obtained by knowledge graph reasoning. It also generates a candidate recommendation list based on the similarity calculation between the personnel ability feature vector and the project requirement feature vector and multi-objective optimization sorting. Step S5: During project execution, the dynamic scheduling module monitors project progress, personnel attendance, and work output indicators in real time. When a project change request is detected, a dynamic scheduling plan is generated and personnel are allocated. Step S6: Use the effect prediction module to predict and analyze the project success rate, staff retention rate and customer satisfaction corresponding to the current matching plan, and evaluate the execution effect of the matching plan based on the prediction results. Step S7: After the project is completed, the performance evaluation module is used to comprehensively evaluate the IT personnel's project delivery quality, work attitude, teamwork ability, technical growth and customer feedback, and the performance evaluation results are fed back to the personnel capability profile database for subsequent personnel matching and dynamic scheduling optimization.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: I. Significantly improve matching efficiency: Through automated feature extraction and intelligent matching algorithms, the traditional manual screening process is shortened to minutes, improving matching efficiency by more than ten times.

[0018] II. Achieve dynamic and flexible scheduling: Through real-time monitoring and reinforcement learning algorithms, it can quickly respond to project changes, automatically generate adjustment plans, and achieve dynamic optimization of personnel allocation.

[0019] Third, it provides the ability to predict the effects: The effect prediction model trained on historical data can predict the project effects in the matching stage, providing data support for decision-making and reducing project risks.

[0020] IV. Comprehensive and objective performance evaluation: A multi-dimensional performance evaluation system can comprehensively reflect the work performance of personnel and provide a basis for talent training and motivation.

[0021] V. Promote knowledge accumulation and inheritance: The knowledge graph module can effectively accumulate project experience and solutions, and reduce knowledge loss caused by personnel turnover. Attached Figure Description

[0022] Figure 1 This application provides a schematic diagram of the system modules. Figure 2 A flowchart illustrating the method steps provided in this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] refer to Figure 1 This invention provides an AI-based intelligent matching and dynamic scheduling module for IT personnel outsourcing, comprising the following modules: The data acquisition module is used to obtain IT personnel data, project requirement data, and business process data from multiple business systems, and to perform unified aggregation and standardized processing on the collected data. The feature engineering module is used to perform structured processing and feature vector construction on the collected data. It generates unified personnel capability feature vectors and project requirement feature vectors through text feature extraction, numerical feature processing, and category feature encoding. The intelligent matching engine module is used to screen, match, sort, and generate a recommendation list of candidates based on project requirements and personnel capabilities. The dynamic scheduling module is used to monitor project progress and personnel status in real time during project execution, and generate dynamic adjustment plans and execute personnel scheduling when changes are detected. The effect prediction module is used to predict and analyze the project success rate, staff retention rate and customer satisfaction after the matching plan is implemented, in order to evaluate the feasibility of the current matching plan. The performance evaluation module is used to evaluate the comprehensive performance of IT personnel based on project execution results and multi-dimensional evaluation data, and to generate performance evaluation results for updating personnel capability profiles. The knowledge graph module is used to build a knowledge graph in the field of IT personnel outsourcing. By associating and modeling personnel skill information, project experience information, technical difficulty information, and solution information, it forms personnel skill graphs, project experience graphs, technical difficulty graphs, and solution graphs. These graphs are then stored and analyzed through a graph database to provide knowledge support for personnel matching, scheduling decisions, and effect prediction.

[0025] In practical applications, an IT outsourcing service company needs to manage its internal personnel resources and project requirements in a unified manner. The system first uses a data acquisition module to uniformly acquire data scattered across multiple business systems. Specifically, the data source access submodule establishes data interface connections with the company's internal human resources system, project management system, customer feedback system, attendance system, and training system, respectively, to obtain basic personnel information, project contract information, project execution data, customer evaluation information, and training record information from each system. Subsequently, the data synchronization submodule executes data synchronization tasks according to a preset cycle, updating newly added or changed data through incremental synchronization, while performing full backups regularly. The consistency of data updates is ensured through timestamp and version identification mechanisms. The data preprocessing submodule further cleans, deduplicates, converts formats, and fills in missing values ​​on the collected data, thereby forming a set of structured and unstructured data, which is then uniformly stored in a data warehouse to provide basic data support for subsequent analysis and processing.

[0026] After data collection, the system uses the feature engineering module to extract and construct features from the raw data. The text feature extraction submodule first performs semantic analysis on unstructured data such as candidate resumes, historical project reports, and technical documents, and extracts text semantic vectors based on the BERT pre-trained language model to generate text features reflecting personnel's technical capabilities and project experience. Subsequently, the numerical feature processing submodule normalizes and bins numerical data such as years of work experience, number of projects completed, and customer ratings to reduce the influence between data of different dimensions. The categorical feature encoding submodule performs one-hot encoding on categorical data such as technology stack categories, job categories, and project types to generate discrete feature representations that can be used for model training. At the same time, the time-series feature extraction submodule extracts personnel capability change trend features based on personnel's historical project participation records and skill growth trajectory using the sliding window method, and dynamically updates personnel capability feature vectors to form a complete personnel capability profile and project requirement feature representation.

[0027] After feature construction is completed, the system uses a knowledge graph module to structurally model the knowledge in the IT personnel outsourcing field. The personnel skills graph sub-module uses IT personnel as the core node, recording the technical skills, skill levels, and skill development trajectories of personnel, and establishing the relationship between personnel and technical skills. The project experience graph sub-module records project information participated in by personnel, including project type, technology stack structure, personnel responsibilities and roles, and project results, and establishes a relationship mapping between personnel and projects. The technical difficulty graph sub-module records the technical difficulties encountered during project implementation and establishes the relationship between technical difficulties and technical fields, project types, and related technology stacks. The solution graph sub-module records solutions and best practice experiences for various technical difficulties and establishes the correspondence between solutions and technical difficulties. The above knowledge graphs are stored in the Neo4j graph database, and knowledge nodes and relationship edges are constructed through a combination of automatic information extraction and manual review, thereby supporting personnel matching, technical reasoning, and knowledge retrieval.

[0028] After obtaining personnel characteristic data and knowledge graph information, the system uses an intelligent matching engine module to achieve intelligent matching between personnel and project requirements. For example, when a company needs to allocate 5 backend development engineers for a 6-month software development project, the system first uses a candidate screening submodule to conduct preliminary screening of candidates based on technical stack requirements, years of work experience, educational requirements, and certification requirements. It then uses an inverted index structure to quickly locate the set of candidates that meet the conditions in the candidate database. Subsequently, the ranking optimization submodule calculates the matching degree score by calculating the cosine similarity between the personnel's ability feature vector and the project requirement feature vector. It also constructs a multi-objective optimization model by combining the personnel's current workload index, historical cooperation effect index, and customer historical preference index. The system calculates the comprehensive score of the candidates using a weighted summation model. Finally, the diversity maintenance submodule evaluates the diversity of the ranking results. When it finds that the technical backgrounds of the recommended personnel are too concentrated, it adjusts the recommendation list by introducing differentiated candidates, thereby generating a more reasonable personnel recommendation scheme.

[0029] During project execution, if changes in requirements or anomalies occur, the system responds in real time through the dynamic scheduling module. For example, when a client requests a new functional module during the project implementation phase, the system first uses the anomaly detection submodule to monitor key indicators such as project progress, staff attendance, and work output in real time, and identifies situations where project resources are insufficient. Subsequently, the impact assessment submodule analyzes the new requirement, assesses its impact on project cycle, resource allocation, and technical complexity, and generates corresponding impact scores and risk levels. The solution generation submodule evaluates possible personnel allocation schemes based on reinforcement learning strategies, and selects the optimal scheduling scheme from multiple candidate schemes by comprehensively considering project completion rate, staff utilization rate, and customer satisfaction through a reward function. The automatic execution submodule automatically executes the scheduling scheme after obtaining authorization, including sending allocation notices to relevant personnel, updating project plans, and synchronizing relevant business system data, thereby completing resource adjustments in a short period of time.

[0030] After the personnel matching plan is determined, the system conducts a feasibility analysis on the matching results through the effect prediction module. The project success rate prediction submodule analyzes project characteristics, personnel configuration characteristics, and historical project data based on the multilayer perceptron model and outputs the project success probability. For example, the predicted success rate of a certain project is 86%. The personnel retention rate prediction submodule predicts the probability of personnel continuing to work during the project cycle through the survival analysis model and gives the short-term retention probability and long-term retention probability. The customer satisfaction prediction submodule predicts the customer's satisfaction rating of the project results through the regression model. The system integrates the above prediction results to form a project effect prediction report. When the predicted success rate is lower than the set threshold, the system automatically prompts to rematch personnel or adjust the configuration plan.

[0031] After the project concludes, the system uses a performance evaluation module to conduct a comprehensive performance evaluation of the IT personnel involved. The performance indicator construction submodule establishes a multi-dimensional performance indicator system that includes project delivery quality, work attitude, teamwork, technical growth, and customer feedback. The evaluation data collection submodule collects evaluation data from project managers, team members, and customers, and normalizes it using a unified scoring standard. The performance analysis submodule uses a 360-degree evaluation method to comprehensively calculate the scores for each dimension and generates a performance radar chart and performance details table. For example, if an engineer's project scores are: project delivery quality 85, work attitude 90, teamwork 88, technical growth 80, and customer feedback 82, the system comprehensively evaluates their performance level as A and updates this result to the personnel capability profile database, providing a reference for subsequent project matching and personnel scheduling.

[0032] In addition, during project execution, when the team encounters complex technical problems, the project manager can also conduct technical queries through the knowledge graph module. The system uses the technical difficulty graph and solution graph to search for similar technical problems and corresponding solutions that have occurred in historical projects, and combines the project experience graph to provide relevant personnel or project cases, thereby helping the project team to quickly locate problem-solving ideas and improve project development efficiency.

[0033] refer to Figure 2 This invention provides an AI-based intelligent matching and dynamic scheduling method for IT personnel outsourcing, comprising the following steps: Step S1: Collect basic information of IT personnel, project requirements, project execution data and customer evaluation data from the human resources system, project management system, customer feedback system, attendance system and training system through the data acquisition module, and store the data in a unified data warehouse after data cleaning, format unification, missing value filling and standardization. Step S2: The feature engineering module performs feature extraction and feature encoding on the collected data. Semantic vectors are extracted from resume text, project reports and technical documents. Work experience years, number of completed projects and customer rating data are normalized. Technology stack category, job category and project type are encoded and corresponding personnel ability feature vectors and project requirement feature vectors are generated. Step S3: Extract knowledge and construct relationships from IT personnel skill information, project experience information, technical difficulty information, and solution information through the knowledge graph module to form a knowledge network that includes personnel skill graph, project experience graph, technical difficulty graph, and solution graph. Obtain personnel skill association information and project experience association information based on graph query and relational reasoning. Step S4: The intelligent matching engine module filters candidates based on the project requirement feature vector and the personnel skill association features obtained by knowledge graph reasoning. It also generates a candidate recommendation list based on the similarity calculation between the personnel ability feature vector and the project requirement feature vector and multi-objective optimization sorting. Step S5: During project execution, the dynamic scheduling module monitors project progress, personnel attendance, and work output indicators in real time. When a project change request is detected, a dynamic scheduling plan is generated and personnel are allocated. Step S6: Use the effect prediction module to predict and analyze the project success rate, staff retention rate and customer satisfaction corresponding to the current matching plan, and evaluate the execution effect of the matching plan based on the prediction results. Step S7: After the project is completed, the performance evaluation module is used to comprehensively evaluate the IT personnel's project delivery quality, work attitude, teamwork ability, technical growth and customer feedback, and the performance evaluation results are fed back to the personnel capability profile database for subsequent personnel matching and dynamic scheduling optimization.

[0034] It should be noted that, unless otherwise specified, the embodiments and features and technical solutions in the present invention can be combined with each other.

[0035] Obviously, the embodiments described above are merely some embodiments of the present invention, not all embodiments. The accompanying drawings show preferred embodiments of the present invention, but do not limit the patent scope of the present invention. The present invention can be implemented in many different forms; rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the patent protection scope of this invention.

Claims

1. An AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing, characterized in that, Includes the following modules: The data acquisition module is used to obtain IT personnel data, project requirement data, and business process data from multiple business systems, and to perform unified aggregation and standardized processing on the collected data. The feature engineering module is used to perform structured processing and feature vector construction on the collected data. It generates unified personnel capability feature vectors and project requirement feature vectors through text feature extraction, numerical feature processing, and category feature encoding. The intelligent matching engine module is used to screen, match, sort, and generate a recommendation list of candidates based on project requirements and personnel capabilities. The dynamic scheduling module is used to monitor project progress and personnel status in real time during project execution, and generate dynamic adjustment plans and execute personnel scheduling when changes are detected. The effect prediction module is used to predict and analyze the project success rate, staff retention rate and customer satisfaction after the matching plan is implemented, in order to evaluate the feasibility of the current matching plan. The performance evaluation module is used to evaluate the comprehensive performance of IT personnel based on project execution results and multi-dimensional evaluation data, and to generate performance evaluation results for updating personnel capability profiles. The knowledge graph module is used to build a knowledge graph in the field of IT personnel outsourcing. By associating and modeling personnel skill information, project experience information, technical difficulty information, and solution information, it forms personnel skill graphs, project experience graphs, technical difficulty graphs, and solution graphs. These graphs are then stored and analyzed through a graph database to provide knowledge support for personnel matching, scheduling decisions, and effect prediction.

2. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The data acquisition module includes: The data source access submodule is used to establish data interface connections with the internal human resources system, project management system, customer feedback system, attendance system, and training system to obtain basic personnel information, project contract information, project execution data, customer evaluation information, and training record information. The data synchronization submodule is used to periodically synchronize and update data in various business systems using a data synchronization strategy that combines incremental synchronization and full backup, and to ensure the timeliness and consistency of data updates through timestamp and version identification mechanisms. The data preprocessing submodule is used to clean, deduplicate, convert formats, and impute missing values ​​in the collected data, thereby forming structured and unstructured datasets. The processed data is then stored in a unified data warehouse for subsequent modules to use.

3. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The feature engineering module includes: The text feature extraction submodule is used to perform semantic analysis on unstructured text data, including resume text, project report and technical document, and extracts semantic vectors based on BERT pre-trained language model to generate text semantic feature vectors. The numerical feature processing submodule is used to perform normalization and binning processing on numerical data, including years of work experience, number of projects completed, and customer ratings, in order to eliminate differences between data of different dimensions. The category feature encoding submodule is used to perform one-hot encoding on categorical data, which includes technology stack categories, job categories and project types, thereby generating discrete feature representations that can be used for model training. The temporal feature extraction submodule is used to extract features from personnel's historical project participation records and skill growth trajectories using a sliding window method to construct personnel capability change trend features and dynamically update personnel feature vectors.

4. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The intelligent matching engine module includes: The candidate screening submodule is used to initially screen candidates based on the technical stack requirements, work experience requirements, education requirements, and certification requirements in the project requirements, and to quickly locate the set of candidates that meet the conditions in the candidate database through an inverted index structure. The ranking optimization submodule is used to comprehensively rank the candidates who have passed the initial screening. It obtains the matching score by calculating the cosine similarity between the personnel's ability feature vector and the project's requirement feature vector. At the same time, it constructs a multi-objective optimization model by combining the personnel's current workload index, historical cooperation effect index, and customer historical preference index, and calculates the candidate's comprehensive score through a weighted summation model. The diversity preservation submodule is used to evaluate the diversity of the ranking results. It judges the concentration of the recommendation results by calculating the feature similarity between recommenders. When the inter-group difference of the recommendation results is lower than a set threshold, the recommendation list is readjusted by introducing differentiated candidates to improve the diversity of the recommendation results.

5. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The dynamic scheduling module includes: The anomaly detection submodule is used to monitor key indicators in real time during project execution, including project progress indicators, personnel attendance indicators, and work output indicators, and to identify abnormal events and trigger early warning mechanisms based on time-series data anomaly detection algorithms. The impact assessment submodule is used to assess the scope of impact of change factors when a project change request is received, and to comprehensively analyze the technical complexity of the change, the time urgency, and the intensity of resource requirements, thereby generating the corresponding impact score and risk level. The scheme generation submodule is used to optimize and search for personnel allocation schemes based on reinforcement learning strategies. It constructs a reward function to comprehensively evaluate project completion rate, personnel utilization rate and customer satisfaction, and selects the optimal scheduling scheme from multiple candidate schemes. The automatic execution submodule is used to execute the scheduling plan after obtaining management authorization, including personnel allocation notification, project plan update and data synchronization update of related business systems, and to monitor the execution status in real time to ensure the stability of the scheduling process.

6. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The effect prediction module includes: The project success rate prediction submodule is used to build a multilayer perceptron prediction model. It takes the matching scheme features as the model input, trains it with historical project data, and outputs the project success probability. The training data includes project features, personnel configuration features, and project result labels. The model is optimized using the cross-entropy loss function and uses dropout technology to prevent overfitting. The employee retention rate prediction submodule is used to predict the probability of employees continuing to work in a project based on a survival analysis model, and to generate short-term retention probability and long-term retention probability by comprehensively analyzing factors such as employee job satisfaction, team cohesion, technical challenges and salary competitiveness. The customer satisfaction prediction submodule is used to build a regression prediction model to predict customer satisfaction ratings for project outcomes. Its inputs include project characteristics, staffing characteristics, and project execution process characteristics, and its output is the predicted satisfaction rating. The three prediction results together form a project performance prediction report. When the project success probability is lower than a preset threshold, the system triggers a re-matching mechanism.

7. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The performance evaluation module includes: The performance indicator construction submodule is used to establish a multi-dimensional performance indicator system, including project delivery quality indicators, work attitude indicators, team collaboration indicators, technical growth indicators, and customer feedback indicators. The evaluation data collection submodule is used to collect self-evaluation data, supervisor evaluation data, peer evaluation data, and customer evaluation data, and to normalize them using a unified scoring standard. The performance analysis submodule is used to comprehensively calculate the scores of each dimension using a 360-degree evaluation method, and generate a performance radar chart and performance details table in a visual way to form a comprehensive performance evaluation result for IT personnel. The evaluation result is then fed back to the personnel capability profile to support subsequent matching and optimization.

8. The AI-based intelligent matching and dynamic scheduling system for IT personnel outsourcing according to claim 1, characterized in that, The knowledge graph module includes: The Personnel Skills Mapping submodule is used to build a personnel skills knowledge network with IT personnel as the core node, record the technical skills, skill levels and skill development trajectories of personnel, and establish the relationship between personnel and technical skills; The Project Experience Graph submodule is used to build an experience knowledge network of personnel participating in projects, record project types, technology stack structure, personnel responsibilities and roles, and project deliverables, and establish a mapping between personnel and projects; The Technical Challenge Map submodule is used to record the technical challenges encountered during project implementation and to establish the relationship between technical challenges and technical fields, project types, and related technology stacks. The Solution Mapping submodule is used to record solutions, handling strategies, and best practice experiences for technical challenges, and to establish the relationship between solutions and technical challenges. The knowledge graph is stored in the Neo4j graph database and knowledge nodes and relationship edges are constructed through a combination of automatic information extraction and manual review to support personnel matching, technical reasoning and knowledge retrieval.

9. A method for intelligent matching and dynamic scheduling of IT personnel outsourcing based on artificial intelligence, characterized in that, The system applied to any one of claims 1 to 8 includes the following steps: Step S1: Collect basic information of IT personnel, project requirements, project execution data and customer evaluation data from the human resources system, project management system, customer feedback system, attendance system and training system through the data acquisition module, and store the data in a unified data warehouse after data cleaning, format unification, missing value filling and standardization. Step S2: The feature engineering module performs feature extraction and feature encoding on the collected data. Semantic vectors are extracted from resume text, project reports and technical documents. Work experience years, number of completed projects and customer rating data are normalized. Technology stack category, job category and project type are encoded and corresponding personnel ability feature vectors and project requirement feature vectors are generated. Step S3: Extract knowledge and construct relationships from IT personnel skill information, project experience information, technical difficulty information, and solution information through the knowledge graph module to form a knowledge network that includes personnel skill graph, project experience graph, technical difficulty graph, and solution graph. Obtain personnel skill association information and project experience association information based on graph query and relational reasoning. Step S4: The intelligent matching engine module filters candidates based on the project requirement feature vector and the personnel skill association features obtained by knowledge graph reasoning. It also generates a candidate recommendation list based on the similarity calculation between the personnel ability feature vector and the project requirement feature vector and multi-objective optimization sorting. Step S5: During project execution, the dynamic scheduling module monitors project progress, personnel attendance, and work output indicators in real time. When a project change request is detected, a dynamic scheduling plan is generated and personnel are allocated. Step S6: Use the effect prediction module to predict and analyze the project success rate, staff retention rate and customer satisfaction corresponding to the current matching plan, and evaluate the execution effect of the matching plan based on the prediction results. Step S7: After the project is completed, the performance evaluation module is used to comprehensively evaluate the IT personnel's project delivery quality, work attitude, teamwork ability, technical growth and customer feedback, and the performance evaluation results are fed back to the personnel capability profile database for subsequent personnel matching and dynamic scheduling optimization.

Images