Jishnu S

Research Assistant

ABB

Mar 2023 - Aug 2023 • 5 mos

Built a Python GUI application orchestrating four MQTT network attack tools (SlowDoS, FloodingDoS, Brute Force, Malformed) integrated from disparate open-source repositories --- eliminating the manual setup overhead that had been bottlenecking the ML detection team's experiment throughput. Implemented YAML-based attack configuration save/load, cutting attack-scenario setup from hours to minutes and enabling reproducible experiments across the project lifetime. Held weekly Agile syncs with the ML detection (blue team) research group, aligning telemetry hooks and validating traces across all simulated attack scenarios --- directly enabling them to train and evaluate detection models on consistent ground truth. Conducted rigorous testing and performance evaluation, measuring attack durations and resource consumption across configurations to surface deployment-time bottlenecks before they hit production research workloads. Documented installation steps and per-tool dependencies, removing the onboarding tax for every future research assistant joining the project.