Hydraulic Modeling & Probabilistic Leak Detection — Water Distribution System — Unalakleet, Alaska

• Developed a Python-based hydraulic modeling and data assimilation system using EPANET for network-wide state estimation and monitoring in a remote water distribution system (4 loops, ~740 population)
• Designed a probabilistic framework for leak detection, localization, and system diagnostics under uncertainty (pipe roughness, demand variability, hydraulic losses)
• Integrated SCADA API data, wireless pressure sensors, and adaptive sampling within an AWS-based data pipeline, enabling continuous monitoring, anomaly detection, and data-driven operational decision-making (e.g., pump scheduling and valve control)
• Built a browser-based digital twin dashboard (Flask + Plotly + Leaflet on AWS EC2) for real-time situational awareness

Real-Time Sensor Quality Control & Data Assimilation — Waller Creek Watershed — Austin, Texas

• Developed a Python-based hydrologic–hydraulic modeling framework using PipeDream solver for unsteady flow simulation and data assimilation in an urban watershed
• Designed an online quality control (QC) algorithm using Extended Kalman Filter (EKF) to detect and correct sensor faults in streaming data, achieving ROC AUC > 0.99
• Developed and deployed a wireless sensing network (4 ultrasonic sensor nodes) to collect continuous water level data for real-time model integration
• Enabled real-time anomaly detection and improved water level prediction, supporting flood alert and monitoring systems