Intrusion Detection System (Government Super App)

Developed deep-learning intrusion detection system for the Indonesian Government Super App, comparing Deep Neural Networks with Multinomial Logistic Regression, Naive Bayes, and AdaBoost.

Focused detection on DoS, Probe, R2L, and U2R attacks by transforming raw logs into analysis-ready inputs.

Implemented and evaluated end-to-end models, ran intrusion simulations, and communicated findings through dashboards.

Achieved 95% detection, deployed analyst interface with Streamlit, and placed 3rd runner-up nationally (TSDN 2023).