Optimizing System Access: A Neural Network Graph Recommender for New Employee Onboarding & Anomaly Detection in Access Permissions

This capstone project explores the development of a neural network graph recommender model for a multinational pharmaceutical and healthcare company to manage system access privileges effectively. The project addresses the challenges of onboarding new employees and the internal mobility of existing staff, which often leads to inefficient and insecure access management. The primary objectives include designing a predictive algorithm to recommend appropriate system access based on roles and historical data, and creating an anomaly detection mechanism to identify security risks. The project leverages Human Resources (HR) and Identity and Access Management (IAM) data to optimize access provision and mitigate security risks, enhancing productivity and cybersecurity within the organization. The scope covers optimizing access management processes, data analysis, and modeling, with evaluation metrics to assess the model’s effectiveness. Limitations include data quality, model generalization, and organizational resistance to change. The project aims to streamline access management and ensure robust cybersecurity measures by integrating advanced data science techniques into the company’s IAM practices.

Objectives of the project:

1. Develop a neural network recommender model.
2. Create a prediction algorithm to predict appropriate system access for new employees.
3. Create an anomaly detection mechanism.
4. Define and implement performance evaluation metrics.
5. Prepare a comprehensive project presentation.