• Researched academic papers and communicated with 3 product security teams to integrated Machine Learning into existing monitoring and logging against SSH brute-force attack abuse case.

• Designed 3 ML-based detection algorithms using LASSO regressor and Decision Tree Classifier.

• Reduced false positive alerts by 70+% and provided coverage for 6 additional API endpoints.

• Monitored AWS logs over 3 weeks through Splunk Enterprise Security and ML Toolkit, identified

5 individual and 10+ distributed attacker suspects.

• Extracted 500+ MB data from OpenStreetMap for 5 US cities, and used handler to generate relational table (.csv) 40,000+ streets from JSON format, with 100+ key-value pairs as features.

• Explored various ML models to predict pedestrian density on streets across 5 cities, achieved validation accuracy as high as 95% (87% when city-blind) with Random Forest Regressor.