2025_EML Capstone Projects

Phased implementation: It played a critical role in incremental change and the rule-based system was a trust-building layer. When this was accepted, machine learning was progressively introduced as an improvement rather than a complete system replacement. In such an organization, change has never been an easy process to implement. Adopting phases provided extra time to validate system behavior step by step and realize how this can enhance the organization in a positive way. Stakeholder engagement: Compliance and security implementations were moved from the design phase. Stakeholders provided feedback on dashboard UI and UX which helped increase system usability. Inclusion created a sense of ownership that turned potential critics into champions of the system. Transparency and explainability: Prioritized using visual tools to be more transparent and explain how decisions were made. Also, alerts were documented to explain features and combinations such as download size, time of login activities as well as privileges which eventually equal to anomaly and captured by the detection system. Logs were collected clearly and integrated for each flagged user activity to help with internal audits to reduce black-box concerns. Adaptability for future development: This system has been built to thrive, and new features can always be added to improve while machine learning models can be restrained. Without starting all over, if flexibility is practiced, the system can keep a stable pace with evolving insider threat vectors. Monitoring and Execution: Tools used for monitoring throughout the implementation phases include Jupiter notebooks and GitHub for code tracking, collaborative updates, and version control. Also, for log auditing, each anomaly detection was paired with a traceable event log that captures raw inputs and model decision details. To understand system outputs, and supporting non technical stakeholders, flask web dashboard is used. Effectiveness of strategies: Project progress was captured by creating weekly update logs, and blockers were identified with collected logs. The documentation strategy ensured that changes were reflected in real time. The bridge between technical and compliance parts of the project was gaped with outputs and visualizations. Additionally, documentation and audit logs facilitated internal communication and accountability, in addition to visual dashboards that significantly enhanced system transparency. Lastly, real-time collaboration and milestone tracking were enabled by Git-based version control.

Back to Table of Contents